Globe Savings Bank, F.S.B. v. United States

65 Fed. Cl. 330, 2005 U.S. Claims LEXIS 124, 2005 WL 1023484

• Court: United States Court of Federal Claims
• Decided: April 29, 2005
• Docket: No. 91-1550C
• Status: Published

Opinion

OPINION AND ORDER

LETTOW, Judge.

INTRODUCTION

In this Winstar-related case,¹ plaintiff Globe Savings Bank, F.S.B. (“Globe”) was a thrift savings bank that operated in Oklahoma in the late 1980s and early 1990s. Phoenix Capital Group, Inc. (“Phoenix”) was the holding company for all of the stock of Globe. The enactment of the Financial Institutions Reform, Recovery, and Enforcement Act (“FIRREA”) on August 9, 1989, Pub.L. No. 101-73, 103 Stat. 183 (codified in scattered sections of Title 12 of the U.S.Code, including 12 U.S.C. § 1464), removed the regulatory underpinnings for Globe’s capital structure and caused it to embark upon and complete a voluntary liquidation of its banking operations.

Globe and Phoenix filed suit in this court on October 29,1991, seeking damages for the government’s breach of a contract to treat capital credits and supervisory goodwill as regulatory capital.² The case was stayed for a substantial period while Wmstar and the other early Winstar-related cases were tried, decided, and resolved on appeal. After the case became active, discovery and pre-trial proceedings were held that established the government’s liability. See Globe Sav. Bank, F.S.B. v. United States, 55 Fed.Cl. 247 (2003) (granting plaintiffs’ motion for summary judgment on liability and denying defendant’s cross-motion). Subsequent proceedings regarding damages led to a partial grant of defendant’s motion for summary judgment and a remission of further issues concerning damages to trial. See Globe Sav. Bank, F.S.B. v. United States, 59 Fed.Cl. 86 (2003).

A nineteen-day trial on damages in the case was held commencing on July 12, 2004 and concluding on August 5, 2004. The first five days of trial were held in Kansas City, Kansas, and a further 14 days of trial were held in Washington, D.C. Post-trial briefs were filed in September and October 2004, and post-trial argument was held thereafter. For the reasons set out in the opinion which follows, the court awards Globe damages for the government’s breach of its contract. Because one further calculation regarding damages must be made by the parties, as explained below, the exact amount of damages to be awarded in a judgment will await further proceedings in the case.

FACTS³

A. Globe’s Inception

1. A failed thrift in an economically depressed geographic area.

OK Federal Savings and Loan Association (“OK Federal”) was a mutual association located in El Reno, Oklahoma that had become insolvent in 1984. PX 76 at 1-2 (“Issues Memorandum” from B. Neuberger to the Federal Home Loan Bank Board (July 9, 1987)); PX 69 at 8 (“S Memorandum” from K. Mowbray to J. Sconyers (June 29, 1987)). As the insurer of OK Federal’s deposits, the Federal Savings and Loan Insurance Corporation (“FSLIC”) sought to reduce its own liability by restructuring the thrift. On April 18, 1985, OK Federal entered into a consent agreement with the Federal Home Loan Bank Board (“Bank Board” or “FHLBB”), allowing the regulators to seek an acquiror or merger partner for the failed thrift and to replace its management, and limiting the operations OK Federal could conduct without prior approval. PX 69 at 8; PX 76 at 2. Two weeks later, the case was transferred to FSLIC for resolution. PX 76 at 2; PX 69 at 1.

For more than a year, neither the Bank Board nor FSLIC received any acceptable bids for OK Federal. PX 76 at 2; PX 69 at 1. During the mid-late 1980s, Oklahoma was economically depressed, and no local banking institution was willing to take on OK Federal’s relatively large accumulated deficits. PX 42 at 3-5 (Business Plan for Globe (Oct. 24, 1986)); PX 69 at 2 (“S” Memorandum); Tr. 2236:10-13 (Test, of Janet Tasker, an official with FSLIC’s Mergers and Acquisitions Division).

2. An unusual acquisition proposal.

One bidder for OK Federal finally emerged. On July 31, 1986, Gerald O’Shaughnessy, an investor from Wichita, and the Phoenix Capital Group, Inc. (“Phoenix”), a Delaware holding company he organized to acquire OK Federal, submitted a proposal to acquire OK Federal with assistance from FSLIC. PX 30 (Acquisition Proposal for OK Federal). Mr. O’Shaughnessy proposed that OK Federal would convert to a stock corporation, that Phoenix would purchase all of the newly issued stock for $3 million in cash, that FSLIC would contribute cash to offset OK Federal’s net worth deficit and provide other indemnifications, and that the Bank Board would provide regulatory forbearances. Id. at 5-16. The proposed forbearances included treating FSLIC’s capital contribution and the supervisory goodwill to be generated in the acquisition through “push-down” accounting as direct additions to the resulting institution’s regulatory net worth. Id. at 13-14. Goodwill would be amortized over a 25-year period by the straight-line method. Id. at 14.

No other prospective acquirer for OK Federal emerged, and the Federal Home Loan Bank of Topeka (“FHLB-Topeka”) considered that “[i]t is highly unlikely that any other acquirer will come to the table with interest in OK Federal.” PX 68 (Mem. from T. Thompson to R. Brick (June 26, 1987)). Economically, Mr. O’Shaughnessy’s proposal was generally acceptable to FSLIC, but the parties engaged in lengthy negotiations over the proposal, focusing primarily on the components of Phoenix’s business plan. Tr. 2862:18 to 2863:11 (Test, of O’Shaughnessy). In the first business plan submitted on October 24, 1986, Phoenix provided a very detailed, definite, and unusual strategy. See PX 42. From the outset, Phoenix planned to use the new entity as a platform for a risk-controlled arbitrage program, pursuant to which the capital credit and supervisory goodwill would be highly leveraged. On the asset side of the ledger, Globe would invest almost entirely in mortgage-backed securities funded by wholesale borrowings on the liability side. Id. at 12-23. The plan sought to limit credit risk to a minimum because the great bulk of the mortgage-backed securities would be insured by the federal government and would consist of pools of geographically diverse loans. Tr. 181:12-24 (Test, of W. Douglas Williams, Globe’s chief executive officer),- 2842:24 to 2843:14 (Test, of O’Shaugh-nessy); PX 69 at 11-12 (“S Memorandum” from K. Mowbray to Jeff Sconyers (June 29, 1987)) (“The credit risk associated with this type of activity is minimal.”); PX 254 at 8 (Investment and Interesb-Rate Risk Management Examination of Globe by J. Curry (Mar. 28,1989)) (“Generally, Globe purchases investments with little or no credit risk, including primarily assets backed by Government National Mortgage Association (GNMA), Federal National Mortgage Association (FNMA), or Federal Home Loan Mortgage Corporation (FHLMC) MBS collateral.”); Id. at 14. On the liability side, Phoenix proposed to employ various hedging techniques to manage interest-rate risk by “duration matching” assets with the liabilities used to fund those assets. PX 42 at 18-19, 36-51; Tr. 194:20 to 195:22 (Test, of Williams). Phoenix hoped to generate stable income at relatively small spreads on mortgage-backed securities in its portfolio, which would be sufficient, given the high leverage proposed and the net operating loss carryforwards (“NOLs”) that OK Federal had produced, both to generate a reasonable level of profits and to shelter those profits from taxation for some years. Tr. 196:19 to 197:13 (Test, of Williams); Tr. 2844:17 to 2845:4 (Test, of O’Shaughnessy). Phoenix planned to employ this risk-controlled arbitrage plan for a transitional period during which it would acquire what it believed to be undervalued deposits and expand its retail customer base in Oklahoma in anticipation of the eventual recovery of the economy there. At that point, it would operate as a more typical banking institution, making secured loans at higher spreads. Tr. 178:24 to 179:4, 198:20 to 199:3, 340:12 to 341:2 (Test, of Williams).

3. Lengthy evaluation of, and negotiations over, the proposal.

To help with its business plan, Phoenix had retained the investment advisory firm of Smith Breeden Associates. That firm provided a detailed briefing to the regulators about the plan in December 1986. PX 47 at 1 (Mem. from J. Arendes to R. Brick (Jan. 29, 1987)). FSLIC and the Bank Board cautiously examined Phoenix’s proposal. Ms. Janet Tasker had been with FSLIC’s Mergers and Acquisitions Division at the time, and she testified that this was “the first time FSLIC had done a [transaction involving] a risk-controlled arbitrage [plan of operation], and it was ... a new concept. It involved a high level of growth.” Tr. 2212:9-12 (Test, of Tasker). In an initial viability analysis performed on February 20, 1987, the Bank Board’s Analysis and Evaluation Division determined that Phoenix’s proposed new entity could be viable, assuming it employed both a high level of expertise and proper hedging techniques in implementing its strategy. PX 52 at 1 (Mem. from G. Schlaseman to J. Tasker & R. Brick (Feb. 20, 1987)). However, Edward Hjerpe of the Office of Policy and Economic Research opined that a leverage multiple greater than 6 to 1 was inappropriate in light of the risks involved with a risk-controlled arbitrage program and that Phoenix should not be allowed to “grow by 500 percent in one year.” PX 48 at 1-2 (Mem. from E. Hjerpe to R. Brick (Feb. 2, 1987)). FSLIC and the Bank Board remained interested in Phoenix’s proposals, and a negotiating session was held on May 7, 1987 between representatives of Phoenix and regulators including Mr. Hjerpe, Ms. Tasker of FSLIC’s Mergers and Acquisitions Division, and representatives of the Analysis and Evaluation Division. PX 61 (FHLBB Mem. to file (May 14, 1987)). Following that meeting, Mr. O’Shaughnessy submitted pro forma financial statements and revised footnotes to comply with Mr. Hjerpe’s guidelines and to answer the regulators’ questions. PX 62 (Letter from G. O’Shaughnessy to R. Sahadi (May 28, 1987)). The Office of Policy and Economic Research and the Principal Supervisory Agent (“PSA”) of FHLB-Topeka found these projections to be reasonable and did not object to Phoenix’s proposed strategy. PX 63 (Mem. from R. Sahadi & E. Hjerpe to R. Brick & J. Tasker (June 1987)); PX 64 (Letter from K. Mowbray to R. Brick (June 11,1987)).

In an updated viability analysis conducted on June 18, 1987, the Bank Board’s Analysis and Evaluation Division again found that Phoenix would be viable, based on the same assumptions employed in its earlier analysis. PX 66 (Mem. from G. Schlaseman & S. Davidson to J. Tasker (June 18, 1987)). Staff at FHLB-Topeka, however, remained concerned that the proposed entity’s very substantial growth would be based on forbearance, rather than tangible, capital and recommended conditioning approval of the acquisition on the new thrift’s maintenance of regulatory capital greater than 6% and on other requirements. PX 69 at 12-13 (“S Memorandum” from K. Mowbray to J. Sco-nyers (June 29, 1987)); Tr. 2234:20 to 2235:10 (Test, of Tasker). FHLB-Topeka opined that assuming these conditions were followed, Phoenix’s risk-controlled arbitrage program would “be a prudent strategy” and recommended that the Bank Board approve Phoenix’s proposal. PX 69 at 12. FSLIC likewise recommended approval, explaining that “[t]he acquirer has satisfied Supervisory and OFSLIC concerns that they have the necessary expertise to initiate and manage such a program.” PX 73 at WOQ413 0278 (“‘A’ Memorandum” from B. Neuberger to FHLBB (July 9, 1987)); see also PX 76 at 7 (“Issues Memorandum” from B. Neuberger to FHLBB (July 9, 1987)) (same); Tr. 2238:4-18 (Test, of Tasker).

4. Consummation of contracts.

On July 22, 1987, the Bank Board approved both the conversion of OK Federal into a stock company and the acquisition of that company by Phoenix. PX 82 (FHLBB Resolution No. 87-793 (July 22, 1987)). The terms of Phoenix’s contractual arrangement with the Bank Board are set out in the accompanying agreements. Pursuant to the deal, OK Federal converted into a stock company named the Globe Savings Bank, F.S.B. (“Globe”), and Phoenix purchased all of Globe’s common stock in exchange for $3 million in cash and a $3 million irrevocable letter of credit. PX 95 (Stock Purchase Agreement (July 31, 1987)); PX 91 (Operating Agreement (July 31, 1987)); PX 93 at 4 (Regulatory Capital Maintenance Agreement (July 31,1987)). FSLIC provided cash assistance of approximately $54.8 million and agreed that Globe could (1) use push-down accounting to reflect the acquisition, (2) amortize the resulting goodwill over the projected life (approximately ten years) of the acquired assets (principally mortgage loans), (3) amortize the resulting capital credit over twenty-five years, and (4) treat both sources of capital as regulatory capital. PX 306 at 1 (Federal Deposit Insurance Corporation Report of Examination (Oct. 19, 1989)); PX 82 at 7 (FHLBB Resolution No. 87-793 (July 22,1987)); PX 85 at 2 (FHLBB Forbearance Letter from J. Buckley to Globe (July 30, 1987)); PX 87 at 24 (Assistance Agreement (July 31, 1987)). For its part, Globe agreed to maintain a minimum capital ratio of 6% for five years and 5% thereafter. PX 93 at 3 (Regulatory Capital Maintenance Agreement (July 31, 1987)). Pursuant to these terms, Globe ultimately included in its regulatory capital the cash contribution of $54.8 million made by FSLIC and approximately $6.8 million of goodwill generated through push-down accounting. PX 306 at 1 (FDIC Report of Examination (Oct. 19,1989)).

B. Globe’s Successful Operations

The Operating Agreement required Globe to submit a revised business plan to provide additional information as requested by the Principal Supervisory Agent of FHLB-Tope-ka or his designee. PX 91 at GB-P-004305. On September 30, 1987, Mr. O’Shaughnessy complied with this requirement. PX 113 (Cover Letter from G. O’Shaughnessy to K. Mowbray); DX 203 (Revised Business Plan (Sept. 30, 1987)). FHLB-Topeka requested further detail with respect to several items of that business plan, PX 120 (Letter from R. Renz to W. Williams (Nov. 3, 1987)), and Globe provided that information the following month. DX 204 (Business Plan Supplement (Dec. 1, 1987)). By letter dated March 18, 1988, FHLB-Topeka approved Globe’s revised business plan. PX 149 (Letter from R. Renz to W. Williams (Mar. 18,1988)).

As a concomitant requirement to adhere to its approved business plan, Globe was obliged to comply with detailed reporting requirements that went well beyond those otherwise applicable. FHLB-Topeka’s approval of the business plan reiterated that the Operating Agreement required Globe to submit to FHLB-Topeka a quarterly report comparing its actual operating results with the results projected in its business plan. Id.; see also PX 91 at GB-P-004305; Tr. 284:9 to 285:2, 314:7-13, 2117:7-21, 2118:7-19 (Test, of Williams). In the Operating Agreement, the Bank Board explained that it added this reporting requirement because it “believe[d] that formulation of and adherence to the business plan is essential to insure the safe and sound operation of [Globe].” PX 91 at GB-P-004304; see also Tr. 285:3-12 (Test, of Williams). In addition to the quarterly report mandated by the Operating Agreement, the approval letter further requested that Globe submit monthly reports of Globe’s investment committee and the minutes from the meetings of that committee. PX 149; Tr. 314:14 to 315:24, 2118:20 to 2119:7 (Test, of Williams). Mr. W. Douglas Williams, the chief executive officer of Globe, testified at trial that pursuant to this requirement, Globe “sent to [FHLB-Topeka] on a monthly basis a very comprehensive, detailed analysis of every single transaction that we were ... transacting. And [FHLB-Topeka] would review it and call us and discuss it with us on a ... monthly basis.” Tr. 315:16-20; see also PX 296 at 18 (FHLB-Topeka’s Regulatory Plan Supervisory Profile (Sept. 22, 1989)).

During the initial months of operation, Globe focused its efforts on resolving the most serious problems with the situation it had inherited. Tr. 116:22 to 117:13 (Test, of Williams). First, OK Federal had on its books “jumbo” CDs, which it had obtained by paying whatever rate was necessary to fund its negative operations. Tr. 318:8 to 321:4, 323:22 to 324:16 (Test, of Williams). Globe adopted a disciplined pricing approach involving a restructuring of the interest rates on deposits based on market rates, which resulted in a run-off of jumbo deposits. Tr. 318:22 to 319:5, 323:2-21, 341:3 to 342:5, 906:1-19, 962:14 to 964:2 (Test, of Williams). Second, Globe inherited approximately $50 to $60 million of nonperforming loans from OK Federal, and those loans showed scant prospects for rehabilitation given the deeply depressed Oklahoma economy at the time. Globe sought to sell those non-performing assets as soon as prudently possible. Tr. 317:2 to 318:7 (Test, of Williams). Disposition by Globe of these non-earning assets benefitted Globe in a way beyond merely converting assets that were not generating earnings into those that could do so. Globe also received a supplemental compensatory payment from FSLIC. Section 3(a) of the Assistance Agreement required FSLIC to pay Globe the amount representing the negative regulatory capital of OK Federal plus $5 million. PX 87 at 22-23. This amount was supposed to have been paid within five days of Globe’s acquisition, id., but FSLIC only partially satisfied its obligation. Globe sought additional compensation from FSLIC, see Tr. 1405:20 to 1408:5 (Test, of Williams); PX 240 (Letter from R. Sommers of FHLBB to W. Williams (Feb. 1, 1989)); PX 241 (Letter from W. Battey of Deloitte Haskins-Sells to W. Williams (Feb. 2,1989)), and on February 1, 1989, Globe reached an agreement with FSLIC pursuant to which FSLIC would pay Globe additional consideration. PX 240, attachment (Mutual Release and Satisfaction Agreement). The total amount Globe eventually received from FSLIC was $58.639 million. PX 240.

By the second quarter of 1988, Globe had addressed the most compelling operating issues and was able to begin implementing its risk-controlled arbitrage program. Tr. 328:13-21, 342:6-13 (Test, of Williams). To provide a foundation for the execution of that program, Globe took steps to expand its retail deposit base. See Tr. 334:8-24, 339:11— 23 (Test, of Williams), 2874:10 to 2875:10 (Test, of O’Shaughnessy). In April 1988, Globe purchased from FSLIC $143.5 million of deposit accounts and five branches of First Federal Savings and Loan Association of Shawnee, Oklahoma (“Shawnee”), an insolvent thrift. PX 185 at 11 (Globe’s consolidated financial statements for period ending June 30, 1988); Tr. 329:11 to 330:3 (Test, of Williams), 2869:23 to 2870:12 (Test, of O’Shaughnessy). Globe “paid” FSLIC a deposit premium of 2.61 percent, or approximately $3.7 million, by accepting the $143.5 million of deposits in return for $139.8 million of cash from FSLIC. PX 170 at 9 (Transfer Agreement between FSLIC and Globe (Apr. 8, 1988)); PX 185 at 11; Tr. 330:4-6, 330:17-20 (Test, of Williams), 2873:14-17 (Test, of O’Shaughnessy). Globe accounted for its purchase premium of $3.7 million as an intangible asset to be amortized over twelve years. PX 185 at 11. Because the $3.7 million premium was an intangible asset, the Shawnee transaction reduced Globe’s tangible capital by the same amount, from negative $6.9 million to negative $10.6 million. Tr. 332:1 to 333:7 (Test, of Williams). Prior to accepting Globe’s bid for Shawnee, FSLIC and FHLB-Topeka had assessed whether Globe’s regulatory capital could support the additional deposits. PX 161 (Executive Summary from J. Buckley to D. Dochow (Apr. 6, 1988)). The Executive Summary recommending approval to the Bank Board stated: “The management of Globe, in the [FHLB] of Topeka’s opinion, is adequate and has expertise in risk arbitrage such that Globe will become and continue to be profitable.” Id. at 3; see also PX 158 (Mem. from S. Root to FHLBB (Apr. 5, 1988)) (recommending approval of Globe’s bid).

With the acquisition of the Shawnee deposits, Globe had a total of approximately $230 million of deposits, which would support about $1.6 billion of wholesale borrowings. Tr. 334:25 to 335:11 (Test, of Williams). The final revision to Globe’s business plan after that acquisition, however, limited growth to about $960 million during the first three years of operation. Tr. 373:3 to 374:5, 377:21 to 379:13 (Test, of Williams).⁴ As a further preparation for implementation of its risk-controlled arbitrage plan, Globe negotiated lines of credit with Wall Street brokerage firms through reverse repurchase agreements and at no standby cost. Tr. 335:12 to 339:8 (Test, of Williams), 2875:14 to 2876:22 (Test, of O’Shaughnessy). These arrangements enabled Globe to purchase mortgage-backed securities in a matter of minutes; the mortgage-backed securities acquired would serve as security for associated borrowings under the repurchase agreements. Tr. 338:15 to 339:4 (Test, of Williams), 2882:6-24 (Test, of O’Shaughnessy).

With this foundation in place, Globe began its initial arbitrage activity. Globe’s first steps included purchasing deeply discounted fixed-rate mortgage-backed securities matched with fixed-rate borrowings. Tr. 349:11-19 (Test, of Williams). Additionally, Globe bought call options to offset prepayment options that came with those fixed-rate assets. Tr. 351:3-9 (Test, of Williams). Thereafter, Globe bought collateralized-mort-gage-obligation floaters, as well as derivative mortgage-backed-security products, including principal-only and interest-only securities, as vehicles to secure cheaper options in falling and rising interest-rate markets. Tr. 351:16 to 352:5, 354:5 to 355:3, 357:7-14 (Test, of Williams). The collateralized-mortgage-obligation floaters represented 59% of its total assets, and they proved to be a highly successful investment, producing 69% of Globe’s total income. See PX 386 at GB-P-0011633 (Analysis of the Profitability of Smith Breeden Investment Activity for Globe, Nov. 1987 to June 1990); Tr. 1314:4 to 1318:17 (Test, of Williams).

In addition, beginning in February 1989 Globe began to purchase bonded tax-advantaged residuals and ultimately purchased approximately $75-$80 million of these securities. Tr. 357:15-21; 361:8-9 (Test, of Williams). These instruments were considerably more valuable to Globe than to most other investors. The bonded tax-advantaged residuals initially threw off “phantom” income that significantly exceeded cash income, but they then switched to throwing off phantom losses that exceeded the cash income. Tr. 2378:24 to 2383:7 (Test, of Andrew S. Davidson, an expert witness appearing on behalf of Globe). In effect, during the early years of the life of these financial instruments, Globe planned to use its carryover NOLs to absorb the phantom income, and then later it would “refresh” those NOLs with the subsequent phantom losses. Tr. 357:19 to 360:18 (Test, of Williams). To a more limited extent, Globe also could balance the phantom income and phantom losses from the bonded tax-advantaged residuals by buying them over time and with varying average maturities. Globe calculated that it had enough NOLs to support approximately $275 million of bonded tax-advantaged residuals. Tr. 366:10-17 (Test, of Williams). Because of the specialized nature of these securities, they traded in a smaller market than mortgage-backed securities generally, and they consequently carried a reduced liquidity-

Following the approvals of Globe’s revised business plan and the Shawnee transaction, Globe rapidly grew its assets. Globe engaged in approximately 70 transactions totaling $754.25 milhon of mortgage-backed securities, with most of these transactions taking place between the second quarter of 1988 and the second quarter of 1989. Tr. 367:24 to 370:1 (Test, of Williams); PX 584-10A (Summary of Globe’s Mortgage-Related Investment Purchases (July 17, 2000)); PX 584-8A (Chart of Globe’s Total GAAP Assets from June 1987 to Dec. 1990). Globe was very selective in the marketplace; its regulatory capital gave Globe time to find appropriate investments. Tr. 371:24 to 372:23 (Test, of Williams). Prior to the passage of FIRREA, Globe’s net interest margin initially increased in tandem with its growth in assets and then remained stable. Tr. 395:18 to 396:13 (Test, of Williams); PX 584-14A (Chart of Globe’s Net Interest Income from Sept. 1987 to Mar. 1991); PX 584-13 (Chart of Globe’s Net Interest Margin from June 1987 to June 1990). The corresponding growth in Globe’s liabilities over the same time period was divided into three categories: (1) FHLB advances ($170.8 million), (2) repurchase agreements with Wall Street investment houses ($321.6 million), and (3) deposits ($202.0 million). Tr. 393:19 to 395:8 (Test, of Williams); PX 584-11 (Chart of Globe’s Liability Composition at June 1989).⁵

Globe had to be especially careful in making investments for its portfolio because of its negative tangible-capital position. As noted previously, after the Shawnee acquisition, Globe had negative tangible capital of $10.6 million. See swpra, at 337. It thus faced the task of building a successful portfolio that consisted in effect of $10.6 million less in assets than liabilities. As Mr. Williams put it in his testimony: “When you have an equal amount of assets and an equal amount of liabilities, your spread would be the same as your net interest margin expressed in basis points.” Tr. 601:4-7. However, with more liabilities than assets, the spread would be greater than the net interest margin. And, it is the net interest margin, not the spread, that reflects the actual results on the balance sheet. Tr. 601:8 to 604:12 (Test, of Williams). Moreover, bank examiners would calculate the net interest margin on earning assets as part of their examination, and disclose the results as part of their examination report. Tr. 604:9-12 (Test, of Williams). As Globe produced positive earnings, however, this adverse circumstance would gradually rectify itself, and Globe’s earning assets would approach and then exceed its liabilities, meaning that its net interest margin would gradually approach its spreads and then eventually exceed them. That is what happened as a result of Globe’s operations.

Beginning in January 1989, FHLB-Topeka conducted its regular examination of Globe, covering the period from November 1, 1987 to January 31, 1989. PX 228 (FHLB-Tope-ka Report of Examination commencing Jan. 3, 1989 and concluding May 4, 1989). FHLB-Topeka found “no matters of significant supervisory concern,” PX 276 (Letter from D. Pittman to Phoenix’s Board of Directors (July 6, 1989)), and opined in its Report of Examination that Globe’s “[investment and risk management strategies have resulted in a favorable interest-rate risk sensitivity profile and a high quality investment portfolio.” PX 228 at 4; see also id. at 1. That report drew upon a special examination of Globe’s investment portfolio and related policies conducted by FHLB-Topeka’s Financial Instruments Specialist, Jeffrey Curry. Tr. 385:14-22 (Test, of Williams), 2656:14 to 2659:2, 2674:5-24, 2823:2-21 (Test, of Curry). As Mr. Curry explained in his report, his examination of Globe was very thorough and comprehensive:

To complete the review of Globe’s investment and interest-rate-risk management, this examiner reviewed all investment and hedging activity over the last five quarters, interviewed management on-site to answer questions regarding certain transactions, thoroughly reviewed all related policies, and analyzed management and board of directors’ reports. In addition, this examiner recently visited the offices of Smith Breeden Associates (SBA), Globe’s investment and hedging advisor, in conjunction with another examination.

PX 254 at 4 (Investment and Interest-Rate Risk Management Examination of Globe by J. Curry (“Curry Report”) (Mar. 28, 1989)); see also PX 230 (Mem. from C. Gee to J. Curry (Jan. 3, 1989)) (requesting that Mr. Curry conduct an examination of Globe’s “Smith-Breeden asset/liability programs”); Tr. 380:17 to 382:17 (Test, of Williams), 2656:14 to 2665:1, 2666:25 to 2667:10, 2679:8 to 2680:20 (Test, of Curry) (testifying that he “reviewed all individual investment transactions and individual hedging transactions over that period.”), 2685:3-16 (Test, of Curry).

The Curry Report emphasized that Globe’s policies for minimizing credit and interest-rate risk had produced very favorable results:

Generally, Globe has adopted an integrated investment and interest-rate-risk management framework that emphasizes the minimization of credit and interest-rate risks. More so than most institutions, Globe has taken a very active role in identifying, measuring, monitoring, and offsetting the specific sources of risk to which it is potentially exposed. This has resulted in a very favorable interest-rate risk sensitivity profile at Globe, as well as a high quality investment portfolio composition.

PX 254 at 4; see also id. at 13-14; Tr. 2676:19 to 2679:7 (Test, of Curry). Mr. Curry praised Globe’s management for its discipline in this regard and its investment expertise in general:

The adherence of Globe’s management to a discipline of truly minimal credit and interest-rate-risk exposure in its investment operations is both unusual and commendable in the thrift industry. This examiner is favorably impressed with both the resources and expertise of the investment management personnel and senior management of the institution.

PX 254 at 11; see also id. at 10 (stating that Glen Wells, Globe’s vice president and chief financial officer, and Mr. Williams, “combined with the resources of [Smith Breeden Associates], result in a very capable investment management operation”); Tr. 2667:11 to 2668:7, 2731:19 to 2732:16 (Test, of Curry). In FHLB-Topeka’s Report of Examination concluding May 4, 1989, Globe received a “MACRO” rating of “1,” the highest possible, on its investment management function. Tr. 2823:11-21, 2825:8-12 (Test, of Curry).⁶

Mr. Curry’s conclusions were confirmed by the Federal Deposit Insurance Corporation (“FDIC”) in an examination of Globe conducted on October 19, 1989. PX 306 (FDIC Report of Examination of Globe (Oct. 19, 1986)). The FDIC opined that “[a]etual income results indicate that management is effectively minimizing interest rate risk. The net interest margin has remained relatively stable during the previous four business quarters.” Id. at 4-a-l; see also PX 296 at 20 (FHLB-Topeka’s Regulatory Plan Supervisory Profile (Sept. 22, 1989)) (“The interest rate sensitivity profile of Globe is attractive and management has demonstrated expertise in matching the duration of assets and liabilities with the highest spread to the funding cost on an option adjusted basis.”). The FDIC report emphasized that Globe achieved positive and stable earnings as well as sufficient liquidity through the successful use of hedging techniques. PX 306 at 1-a.

Globe’s growth and Smith Breeden’s role in carrying out Globe’s investment strategy tailed off in June 1989, as Globe became aware that Congress was considering the legislation that became FIRREA. Tr. 391:1— 10 (Test, of Williams). In particular, Globe made no further purchases of bonded tax-advantaged residuals after June 1989. Tr. 361:13 to 363:20 (Test, of Williams). Faced with the early versions of the new legislative proposals, Messrs. O’Shaughnessy and Williams had several meetings with the Bank Board’s management and with federal legislators from Oklahoma and Kansas to request that the legislation grandfather Globe’s contract with the government respecting the regulatory treatment of Globe’s capital credit and supervisory goodwill. Tr. 411:5 to 417:8 (Test, of Williams), 2887:15 to 2890:19 (Test, of O’Shaughnessy); PX 252 at 1-2 (Minutes of Regular Meeting of Globe’s Board of Directors (Mar. 17, 1989)). Alternatively, Globe sought permission to conduct a “voluntary” dissolution of the institution. Tr. 417:9 to 419:4 (Test, of Williams), 2890:20 to 2896:10 (Test, of O’Shaughnessy); PX 252 at 1-2; PX 278 (Letter from P. Barnett to J. Luke & D. Dochow (July 11,1989)).

C. Effects of FIRREA

The enactment of FIRREA on August 9, 1989 elided or phased out the ability of thrifts to count goodwill as capital for regulatory purposes.⁷ The following month, Globe’s legal counsel interpreted FIRREA to have continued to allow capital credits to be included in regulatory capital and to have grandfathered acquisitions completed prior to 1988, but recognized that because the newly-created Office of Thrift Supervision (“OTS”) had not yet issued regulations or interpretative rulings, the meaning and impact of the statute were uncertain. Tr. 2897:13 to 2899:12 (Test, of O’Shaughnessy); PX 294 (Letter from P. Barnett to Globe’s Board of Directors (Sept. 22, 1989)). Mr. O’Shaughnessy wrote the director of OTS, attaching Globe’s counsel’s legal opinion and soliciting OTS’s agreement with that opinion as well as a meeting to further discuss the matter. Tr. 419:16 to 421:12 (Test, of Williams), 2900:21 to 2902:20 (Test, of O’Shaughnessy); PX 302 (Letter from G. O’Shaughnessy to D. Wall (Oct. 5, 1989)). The director of OTS met with Globe, but he did not at that time take a position respecting OTS’s interpretation of FIRREA as it applied to Globe’s capital credit. Tr. 421:6 to 422:1 (Test, of Williams), 2902:12 to 2903:22 (Test, of O’Shaughnessy).

OTS announced its position in the regulations it promulgated pursuant to FIRREA in November 1989. 54 Fed.Reg. 46,845 (Nov. 8, 1989). These regulations indicated in a parenthetical that capital credits, like supervisory goodwill, could no longer be counted toward regulatory capital. See 12 C.F.R. §§ 567.1(a)(2)(iv) and (w) (1990) (“The term ‘qualifying supervisory goodwill’ means, for eligible savings associations: (1) Any unam-ortized goodwill (FSLIC Capital Contributions, as reported in the September 30, 1989 Thrift Financial Report) that existed on April 12, 1989 resulting from prior regulatory accounting practices less any amortization id. § 567.5(a)(2)(iii)(B) (1990). See also 54 Fed.Reg. at 46,860 (“[C]ore capital also includes, until December 31, 1994, goodwill (FSLIC Capital Contributions) resulting from prior regulatory accounting practices. This goodwill is gradually phased out over this five year period.”). Globe’s management referred to this clarifying provision rejecting Globe’s interpretation of FIRREA as “the Globe parenthetical.” Tr. 2903:22 to 2904:17 (Test, of O’Shaughnessy); see also Tr. 428:17 to 429:13 (Test, of Williams).

By December 31,1989, Globe had achieved a positive, but very small, level of tangible capital. See Tr. 623:9-17 (Test, of Williams); PX 584 at 58 (Expert Report of W. Williams (Nov. 5,1999)); PX 584-15 (Graph of Globe’s tangible capital growth from June 1987 to June 1990); PX 584-16 (Graph of Globe’s actual and projected net income from June 1988 to June 1990); DX 122 at 1 (Letter from W. Williams to D. Pittman (Apr. 6, 1990)). Counting the capital credit and supervisory goodwill, Globe’s total regulatory capital in December 1989 was $59.6 million, while its tangible capital was only approximately $261,000. See PX 584-12 (Table of Globe’s historical financial performance). Its operations had been sufficiently profitable that it had had enough income to cover amortization of goodwill and make up its inherited capital deficit. Globe nonetheless relied on its remaining unamortized goodwill to provide the regulatory capital needed to satisfy regulatory standards and to support its risk-controlled arbitrage strategy.

In late 1989 and early 1990, the FDIC, the successor to FSLIC under FIRREA, conducted an examination of Globe. The FDIC did not take into account Globe’s supervisory goodwill or capital credit and thus found that Globe’s capital, although positive, was “grossly inadequate.” PX 306 at 1 (FDIC Report of Examination (Oct. 19, 1989)); Tr. 454:4 to 455:12 (Test, of Williams). For that reason, the FDIC rated Globe a composite “5,” on the MACRO scale, the lowest possible rating, “reserved for institutions with an extremely high immediate or near term probability of failure.” PX 306 at 1-a-l; Tr. 455:13 to 456:2, 1072:2-16, 1075:25 to 1076:21 (Test, of Williams). As the FDIC’s report of examination stated, such a rating reflected the FDIC’s view that

[t]he volume and severity of weaknesses or unsafe and unsound conditions are so critical as to require urgent aid from stockholders or other public or private sources of financial assistance. In the absence of urgent and decisive corrective measures, these situations will likely require liquidating and the payoff of depositors, disbursement of insurance funds to insured depositors, or some form of emergency assistance, merger or acquisition.

PX 306 at 1-a-l. In a letter to Globe attaching that report, the regional director of FDIC, Kenneth Walker, stated that Globe “has an inadequate level of tangible capital,” that “a very significant injection of capital is needed to restore the institution to a viable condition,” and that FDIC “is still evaluating the need to recommend supervisory action.” PX 356 (Letter from K. Walker to Globe’s Board of Directors (Feb. 21, 1990)); Tr. 1070:15 to 1072:1 (Test, of Williams).

Shortly after- receiving the FDIC’s examination report and letter, Globe’s management met with Mir. Walker in Dallas, Texas. Tr. 456:3-13 (Test, of Williams), 2907:7 to 2911:4 (Test, of O’Shaughnessy). At that meeting, Mr. Walker advised Mr. O’Shaughnessy, “don’t take it personally” if the government seized Globe. Tr. 456:14-23 (Test, of Williams), 2912:9 to 2913:3 (Test, of O’Shaughnessy). Thereafter, Globe “began immediately to downsize the institution.” Tr. 456:24 to 458:1 (Test, of Williams), 2913:4 to 2914:20 (Test, of O’Shaughnessy). Less than a month after the meeting in Dallas, in March 1990 the FDIC contacted Globe to schedule a liquidation examination. Tr. 458:7-23 (Test, ofWilliams).

Just as Globe was endeavoring to avoid being seized by the FDIC, the Office of Thrift Supervision also took action regarding Globe. Promptly after OTS issued its regulations implementing FIRREA, OTS-Topeka instructed Globe that because Globe would fail to meet the new capital requirements to become effective December 7, 1989, Globe was required to submit a capital restoration plan by January 8, 1990. PX 321 (Letter from R. Karr to B. Gragert (Nov. 17, 1989)); Tr. 422:8 to 423:22, 1053:11 to 1054:12 (Test, of Williams). In December 1989, OTS-Topeka informed Globe that “a thorough legal and policy review of forbearances is currently being conducted” by OTS, during which time Globe’s capital plan “should not rely on [capital and accounting] forbearances” for meeting capital requirements, and reiterated that Globe should be prepared to submit its capital plan by January 8, 1990. PX 329 (Letter from R. Karr to W. Williams (Dec. 22,1989)); Tr. 425:10-20, 1054:20 to 1055:17 (Test, of Williams).

During this period, OTS-Topeka carried out an intensive internal evaluation of what course of action should be adopted respecting Globe. A meeting was held on January 3, 1990 involving the District Director (Ronald Karr), the Director of Supervision (Louis Roy), the Senior Supervisory Agent (Douglas Pittman), District Counsel (Brian McConnally), Associate Legal Counsel (James Bloss), and a Supervisory Analyst (Diane Fischman) to address Globe’s capital position. PX 343 (Mem. to file by Diane Fischman (Jan. 19, 1990)). The group decided that

although our Office will not require Globe to submit a capital plan until the forbearance issue is resolved, Globe is to comply with the following supervisory directives: 1. Liability growth is not to exceed interest credited; [and] 2. Submission of two five-year pro formas (one which assumes the continuation of all forbearances and the other which assumes the abrogation of all forbearances).

Id. at 1. The officials did not favor a reference to the Resolution Trust Corporation (“RTC”): “In terms of a potential transfer to the RTC, no issues of supervisory concern were noted in the OTS or FDIC ROEs [Reports of Examinations] completed in 1989 other than the issues of capital status.” Id. at 2. They related that “Globe [wa]s performing significantly above business plan projections in terms of net operating income of $2.5 million per year.” Id. As they understood the RTC’s posture, it “[wa]s not interested in liquidating institutions with negative or low tangible capital which report stable earnings and follow relatively low-risk operating strategies; however, these statements ha[d] not been provided in writing.” Id. Finally, the OTS-Topeka officials concluded that Globe “d[id] not appear to pose risk to the SAIF,” the then-new deposit insurance fund for thrifts. Id.

In a subsequent letter dated January 5, 1990, Douglas Pittman, the Senior Supervisory Agent of OTS-Topeka, noted that the conclusions of OTS’s review of forbearances “may have a material impact on the issue of [Globe’s] compliance with the recently promulgated capital regulations.” PX 335 at 1 (Letter from D. Pittman to W. Williams (Jan. 5, 1990)). Pending OTS’s decision whether to breach its contract with Globe, Mr. Pittman directed Globe not to grow beyond the level of net interest credited.. Id.; Tr. 424:5-24 (Test, of Williams). His letter informed Globe that because OTS’s review had not yet been completed, Globe would not be required to submit a capital plan on January 8, 1990. PX 335 at 2. Instead, Mr. Pittman directed Globe to submit five-year pro forma balance sheets and income statements for two alternative scenarios, one projecting Globe’s condition if OTS continued to allow Globe to rely on its capital forbearances, and the other showing the effects on Globe if OTS abrogated those forbearances. PX 335 at 1-2; Tr. 425:21 to 427:23 (Test, of Williams). For both scenarios, Globe was instructed to hold interest-rate relationships constant as of December 31,1989. PX 335 at 2; Tr. 1057:6 to 1058:4 (Test, of Williams).⁸ Mr. Pittman further informed Globe that any questions should be directed to Supervisory Analyst Diane Fischman. PX 335 at 2. Several days after OTS-Topeka provided Globe with this letter, OTS issued guidance on the status of regulatory forbearances in Thrift Bulletin 38-2. That Bulletin provided, in pertinent part, that:

[OTS] is applying the new capital standards to all savings associations, including those associations that have been operating under previously granted capital and accounting forbearances. [FIRREA] eliminates these forbearances. Ml savings associations presently operating with these forbearances, therefore, should eliminate them in determining whether or not they comply with the new minimum regulatory capital standards. (Any FSLIC capital contribution that resulted in the creation of goodwill will be subject to the requirements for goodwill established in the capital regulation).

PX 336 (OTS Thrift Bulletin 38-2 (Jan. 9, 1990)); see also Tr. 1058:24 to 1059:14 (Test, of Williams).

On January 12, 1990, Globe responded to Mr. Pittman’s letter of January 5 by submitting a request for a capital exception. PX 339 (Globe’s Request for Capital Exception (Jan. 12, 1990)); Tr. 429:20 to 433:11, 1058:5-23, 1069:25 to 1070:7 (Test, of Williams). As Globe explained in that request, because OTS’s directives applied to Globe only if it were considered a capital-deficient institution, Globe submitted projections only for a breach scenario in which Globe would shrink and undergo a voluntary solvent liquidation. PX 339 at cover letter & 17-18; Tr. 432:18 to 433:19, 1058:5-23 (Test, of Williams).⁹ Mr. Pittman responded to Globe’s submission by way of a telephone call to Mr. Williams in which Mr. Pittman indicated his displeasure that Globe had submitted projections for only one strategy in the breach scenario and had not provided any information for the non-breach scenario. Tr. 433:20 to 434:19, 488:22 to 489:15 (Test, of Williams). Mr. Pittman followed up on this phone call with a letter dated February 5, 1990, in which he stated, “[a]ll alternatives must be analyzed.” PX 349 at 1 (Letter from D. Pittman to W. Williams (Feb. 5, 1990)); Tr. 490:7-20 (Test, of Williams). In a further discussion by telephone with Mr. Williams, Mr. Pittman clarified that by “all alternatives,” he was interested in five specific scenarios. Tr. 436:24 to 438:6 (Test, of Williams). The first was the non-breach scenario in which the government would allow Globe to continue using its regulatory capital under the terms of the contract. Tr. 438:7-15 (Test, of Williams). The second scenario required Globe to remain at its asset size of December 31, 1989, at which time Globe had $713 million in assets. Tr. 438:24 to 439:15 (Test, of Williams). Because Globe had begun downsizing and reduced its size to approximately $467 million by mid-February 1990, the third scenario was to assume that Globe remained at that lesser size. Tr. 439:16-25 (Test, of Williams). Further, given that Globe was seeking to sell some of its branches in connection with its plans to shrink, Tr. 489:16 to 490:1, 492:15 to 494:4 (Test, of Williams), the fourth scenario assumed that Globe would sell branches in June 1990 and remain at an asset size of approximately $250 million. Tr. 440:1-11 (Test, of Williams). The final projection assumed that the branch sale would occur one year later. Tr. 440:12-22 (Test, of Williams).

As Mr. Pittman had instructed Globe, Mr. Williams and Glen Wells, the chief financial officer of Globe, posed to Diane Fischman their questions seeking further details as to what assumptions they should make in these projections. Tr. 440:23 to 441:13 (Test, of Williams). With respect to interest rates, Ms. Fischman instructed them to use the implied interest rate forecast set out in TB 36-1 through the fifth quarter, see supra, at 342-43 n. 8, but then to keep the rates constant. Tr. 441:14 to 443:6, 1060:1 to 1066:2, 1066:21 to 1069:5 (Test, of Williams); DX 450 at 1 & Table 1. Regarding spread relationships, she requested that Globe use a particular spread that was much less favorable than what was called for by TB 36-1, i.e., the spread relationship that existed at the time. Tr. 443:7 to 444:16, 448:5 to 449:16, 1068:5 to 1069:5 (Test, of Williams); DX 450 at 1-2 & Table 2. Finally, Ms. Fischman instructed Globe to follow the assumptions for prepayment rates specified in TB 36-1. Tr. 444:17 to 445:5, 1066:3-20 (Test, of Williams); DX450at2.¹⁰

On April 6, 1990, Mr. Williams hand delivered to Mr. Pittman the financial projections using the five scenarios Mr. Pittman had requested and applying the assumptions Ms. Fischman had prescribed. DX 122 at WOR637 0125 (Letter from W. Williams to D. Pittman) (“The accompanying financial projections are submitted in response to your [Mr. Pittman’s] letters of January 5, 1990 and February 5, 1990, as well as telephone requests from Diane Fischman.”). Those projections included a scenario in which Globe would sell four of its branches to Equity Bank for Savings, F.A., with which Globe had been involved in negotiations. Id. at 2, Exs. D & F. At a meeting held on April 18, 1990 between representatives of Globe and OTS-Topeka, Globe was informed that OTS had completed its review of regulatory forbearances and had decided to abrogate all such forbearances. Tr. 612:24 to 615:2 (Test, of Williams); PX 373 (Letter from L. Roy to W. Williams (Apr. 23, 1990)) (referencing the meeting held on April 18, 1990). OTS-Topeka confirmed this decision in a letter formally notifying Globe that “FIRREA supersedes the accounting forbearance previously granted to [Globe]; therefore, this forbearance may not be included for the purpose of’ computing regulatory capital. PX 373. Because of OTS’s decision, “Globe [wa]s in noncompliance with its minimum capital requirements for tangible, core, and risk-based capital.” Id.

Because OTS’s interpretation of FIRREA rendered Globe noneompliant with the minimum capital requirements, the April 23 letter also notified Globe that it would have to submit a capital plan within sixty days. Id. Globe submitted its capital plan on June 22, 1990, and OTS-Topeka subsequently approved that plan. PX 405 (Letter from L. Roy to W. Williams (Sept. 25, 1990)); Tr. 1470:15 to 1471:20 (Test, of Williams). As the government recognized, its decision to abrogate Globe’s regulatory forbearances “caused management to downsize the institution.” PX 508 at 1 (FDIC Report of Examination (Apr. 24,1992)); see also PX 485 at 4-5 (OTS Report of Examination (Feb. 4, 1991)). In particular:

In order to meet regulatory capital requirements mandated by the FIRREA, management accomplished significant shrinkage of the balance sheet during the quarter ended December 31, 1990. The branch network; including $160,505,000 in deposits, $11,853,000 in loans, and $1,443,000 in fixed assets; was sold at December 31, 1990. In addition, the MBS [mortgage backed security] and CMO [col-lateralized mortgage obligation] portfolios were greatly reduced and their attendant funding sources retired, reducing total assets to $71,696,000 at December 31, 1990.

PX 435 at 6-7. Respecting Globe’s sale of its branch network, its proposal to sell branches to Equity Bank for Savings, F.A. was denied by OTS. Tr. 516:1 to 517:6, 758:6-16 (Test, of Williams); PX 435 at 5; PX 410 at 1 (Internal Mem. to Globe employees (Oct. 17,1990)) (announcing sale of branches to MidFirst). Thereafter, Globe sold six of its seven branches to MidFirst Bank, S.S.B., Oklahoma City (“MidFirst”), realizing a very small core-deposit premium of 0.7% on the sale. Tr. 754:13 to 763:21; PX 408 (Letter from J. Laisle, president and chief executive officer of MidFirst, to M. Boswell-Carney, OTS (Oct. 16, 1990)) (attaching application to purchase six of Globe’s branches); PX 409 (Purchase and Assumption Agreement between MidFirst and Globe (Oct. 16, 1990)). In December 1990, when Globe’s sale of its branch network to MidFirst closed, Globe sold all but the Harrah branch, which it retained for its home office and as a base from which it might complete its liquidation. Tr. 754:18-20, 755:5-15 (Test, of Williams); PX 410 at 1.

The FDIC applauded Globe’s efforts to shrink drastically in size and achieve capital compliance. In an examination of Globe in 1991, the FDIC opined that Globe had “sold off assets and deposit liabilities in impressive fashion to meet regulatory capital requirements.” PX 454 at WOQ264 1376 (Letters from K. Walker to R. Karr dated June 17, 1991, and to Globe’s Board of Directors dated July 1, 1991, attaching FDIC Report of Examination (Apr. 29, 1991)). Thereafter, by June 30, 1993, Globe had paid off all depositors of the Harrah branch. Tr. 765:18 to 766:14 (Test, of Williams). At that point, Globe relinquished its thrift charter, kept Globe alive as a corporate entity, and paid dividends to, and returned the initial capital contribution of, Phoenix. In total, Globe remitted approximately $7.7 million to Phoenix. Tr. 642:19 to 643:9 (Test, of Williams).

ANALYSIS

Compensation by way of an award of money damages to the promisee is the primary means of redressing a breach of contract. 24 Samuel Williston & Richard A. Lord, A Treatise on the Law of Contracts § 64:1, at 4-5 (4th ed.2002) (hereafter “Williston”). Depending on the circumstances surrounding the breach, three types of contractual damages are potentially available to a promisee: (1) expectancy damages, the amount representing “the benefit of his bargain by being put in as good a position as he would have been in had the contract been performed,” (2) reliance damages, the “reimburse[ment] for loss caused by reliance on the contract by being put in as good a position as he would have been in had the contract not been made,” and (3) restitution damages, the “restoration] to him [of] any benefit that he has conferred on the other party.” Restatement (Second) of Contracts § 344 (1981) (hereafter “Restatement (Second) Contracts”). Expectancy damages ordinarily serve as the basis for an award of contractual damages. LaSalle Talman Bank, F.S.B. v. United States, 317 F.3d 1363, 1371 (Fed.Cir.2003); Restatement (Second) Contracts § 344 cmt. a; Williston § 64:2, at 30.

A plaintiff is entitled to a recovery of expectancy damages upon demonstrating by a preponderance of the evidence that

(1) the loss was the proximate result of the breach; (2) the loss of profits caused by the breach was within the contemplation of the parties because the loss was foreseeable or because the defaulting party had knowledge of special circumstances at the time of contracting; and (3) a sufficient basis exists for estimating the amount of lost profits with reasonable certainty.

La Van v. United States, 382 F.3d 1340, 1351 (2004) (quoting Energy Capital Corp. v. United States, 302 F.3d 1314, 1325 (Fed.Cir. 2002); citing Bluebonnet Sav. Bank, F.S.B. v. United States, 266 F.3d 1348, 1355 (Fed. Cir.2001) (“Expectation damages are recoverable provided they are actually foreseen or reasonably foreseeable, are caused by the breach of the promisor, and are proved with reasonable certainty.”)).

Expectancy damages are often measured by the amount of profits lost but include any other losses caused by the breach, including incidental losses. Glendale Fed. Bank, FSB v. United States, 239 F.3d 1374, 1380 (Fed.Cir.2001) (citing Restatement (Second) Contracts § 347); Williston § 64:2, at 30. “Incidental losses include costs incurred in a reasonable effort, whether successful or not, to avoid loss, as where a party pays brokerage fees in arranging or attempting to arrange a substitute transaction.” Restatement (Second) Contracts § 347 cmt. c; see also Williston § 66:55-56, at 664-677.

A. Causation

Globe has proved by a preponderance of the evidence that the losses it claims as expectancy damages were the proximate result of the government’s breach of contract, i.e., the causal connection between the breach and those losses has been “ ‘definitely established.’ ” California Fed. Bank v. United States, 395 F.3d 1263, 1268 (2005) (internal quotation and citations omitted). As the regulators themselves acknowledged, “[subsequent to the implementation of [FIRREA, Globe’s] capital position was grossly inadequate. This inadequacy caused management to downsize the institution.” PX 508 at 1 (FDIC Report of Examination (Apr. 24, 1992)); see also PX 435 at 4-5 (OTS Report of Examination (Feb. 4, 1991)) (“The abrogation of the forbearances also caused Globe to be out of compliance with the newly-created capital requirements.”). More specifically, OTS’s Report of Examination in February 1991 described succinctly (and accurately) the actions taken by Globe’s management to respond to the breach:

In order to meet regulatory capital requirements mandated by the FIRREA, management accomplished significant shrinkage of the balance sheet during the quarter ended December 31, 1990. The branch network; including $160,505,000 in deposits, $11,853,000 in loans, and $1,443,000 in fixed assets; was sold at December 31, 1990. In addition, the MBS and CMO portfolios were greatly reduced and then1 attendant funding sources retired, reducing total assets to $71,696,000 at December 31,1990.

PX 435 at 6-7; see also id. at 8 (“The asset sales are pursuant to [Globe’s] strategy of balance sheet shrinkage to meet capital requirements mandated by the FIRREA____ Effective December 31, 1990, the institution sold the bulk of its branch network, retail deposits, and performing loans; reducing total assets to $71,696,000.”). OTS-Topeka was acutely aware that the government’s breach caused Globe to operate passively and forego investment opportunities. The Report of Examination in February 1991 recited Globe’s post-shrinkage plans: “President Gragert stated that the institution is in a reactive mode pending changes in the regulatory environment. Management’s intention is to passively operate through the remaining two-year purview of the Operating Agreement. No further asset investments are planned until the CMOs pay down and brokered deposits and reverse repurchase borrowings are retired.” Id. at 8.

The government does not strongly contest this evidence of proximate causation. Instead, the government skips over the steps Globe took to shrink its portfolio and focuses principally on the final steps Globe took to liquidate its thrift banking operations after its sale of the portfolio assets and retirement of attendant borrowings. In this respect, the government draws upon the favorable results from Globe’s sale of the portfolio assets and retirement of attendant borrowings to contend that Globe had come into capital compli-. anee just before it sold its branches to Mid-First. Based upon that circumstance, the government surmises that Globe liquidated not to mitigate the breach, but to “avoid operating in ... the more turbulent interest rate and prepayment environment of the 1990s.” Def.’s Posh-Trial Br. (“Def.’s Br.”) at 12-13. Two of the government’s expert witnesses, W. Barefoot Bankhead and Terry L. Musika, testified to that effect at trial. Tr. 4217:1-25 (Test, of Mr. Bankhead), 4825:12 to 4834:1 (Test, of Mr. Musika). As the government would have it, once Globe completed the disposition of its duration-matched asset-liability portfolio, it could have continued in operation as a small, traditional thrift.

However, the government’s position and the testimony of Messrs. Bankhead and Mu-sika about Globe’s viability after disposition of its asset-liability portfolio and before sale of its branches are contravened by contemporaneous evidence in the trial record. As Globe illustrated in its projections prescribed by OTS-Topeka, Globe could not have remained profitable on a going-forward basis without overcoming its fixed costs of operations by selling its branches:

A reduction in size necessitated by the apparent intent to abrogate forebearanees granted in 1987 results in substantially decreased earnings unless it is accompanied by a deposit transfer allowing the necessary reduction in operating costs. Any delay in implementing such a deposit transfer has a negative impact as it increases the period of time Globe must operate without the spread earning assets necessary to offset operating costs.

DX 122 at WOR637 0127-28 (Letter from W. Williams to D. Pittman (Apr. 6, 1990)) (attaching financial projections); see also id. at WOR637 0248. Notably, the staff at OTS-Topeka acknowledged that the issue facing Globe in early 1990 was whether it could survive as a relatively small institution operating without an arbitraged portfolio. Each of the optional scenarios that staff at OTS-Topeka asked Globe to model contemplated operation of a smaller bank, and two of the four optional scenarios assumed that branch sales would occur. See supra, at 343-44.

In sum, Globe has proved by a preponderance of the evidence that but for the breach, it would not have shrunk to the point of liquidation and that it would have maintained its pre-FIRREA operations. The breach was the definite, direct, and primary cause of Globe’s shrinkage to ultimate liquidation. See California Fed., 395 F.3d at 1267-68.

B. Foreseeability

Unforeseeability constitutes a limitation on damages, and a claimant for damages as a result of a breach of contract bears the burden of showing that its losses were foreseeable. See Home Sav., 399 F.3d at 1355; La Van, 382 F.3d at 1351; Energy Capital, 302 F.3d at 1325; Restatement (Second) Contracts § 351. In this ease, Globe has demonstrated that its losses are “foreseeable as a probable result of a breach because [they] follow[] from the breach [of contract] ... in the ordinary course of events.” Restatement (Second) Contracts § 351(2). That Globe’s damages were foreseeable is evident from the detailed business plan examined by the regulators, and approved by them both in advance of the OK Federal acquisition and thereafter in accord with the Operating Agreement. During the negotiations over Phoenix’s proposed acquisition of OK Federal, the unusual nature of the business plan and its reliance on a risk-controlled arbitrage system to build a large portfolio of mortgage-backed securities matched with borrowings caused the regulators to be wary and skeptical about its viability and to examine the plan closely. After a detailed evaluation of the financial projections in the plan, however, both the Bank Board’s Office of Policy and Economic Research and the PSA of FHLB-Topeka found the projections to be reasonable. See supra, at 334-35 (citing PX 63 (Mem. from R. Sahadi & E. Hjerpe to R. Brick & J. Tasker (June 1987)); PX 64 (Letter from K. Mowbray to R. Brick (June 11, 1987))). Even then, because FHLB-Topeka remained concerned that Globe’s projected significant growth would be based only on forbearances rather than tangible capital, FHLB-Topeka had recommended that approval for Phoenix’s acquisition of Globe should be conditioned on the thrift’s maintenance of capital at a level greater than 6%, among other things. See supra, at 335 (citing PX 69 at 12-13 (an “S Memorandum” from K. Mow-bray to J. Sconyers (June 29, 1987)); Tr. 2234:20-2235:10 (Test, of Tasker)). That recommendation was carried out in the Bank Board’s approval of Phoenix’s acquisition. See supra, at 335 (citing PX 82 (FHLBB Resolution No. 87-793 (July'22, 1987)) (approving the conversion of OK Federal and acquisition of the new entity by Phoenix)). With the change, FHLB-Topeka concluded that, “[g]iven the individuals that the new association will employ, we believe that the proposed business plan can be successfully implemented by [Globe].” PX 64; see also supra at 335.

Additionally, in accord with the Operating Agreement, Globe submitted a revised post-acquisition business plan reflecting additional information requested by FHLB-Topeka, and Globe also submitted further revisions to that plan in connection with the Shawnee transaction and otherwise as it gained experience under the plan. See supra at 336, 337 & n. 4. Those revisions were approved. PX 149 (Letter from R. Renz to W. Williams (Mar. 18, 1988)). Globe was also subject to stringent reporting requirements that called for submission to the regulators of a detailed analysis of every transaction it entered. See supra at 336. In recommending approval of the Shawnee transaction, the Executive Director of the Office of Regulatory Policy, Oversight and Supervision stated that “[t]he management of Globe, in [FHLB-Topeka’s] opinion, is adequate and has expertise in risk arbitrage such that Globe will become and continue to be profitable.” PX 161 at 3.

In short, the regulators were fully aware that Globe’s profits would be based on a capital structure that consisted virtually entirely as an initial matter of regulatory capital comprised of capital credits and supervisory goodwill. See PX 69 at 12 (“S Memorandum” from K. Mowbray to J. Sco-nyers (June 29, 1987)) (noting “the fact that this growth will be supported by RAP [Regulatory Accounting Procedures] capital, and not GAAP capital”). Lost profits and costs incidental to winding down the business were the foreseeable result of disallowing Globe’s use of its capital credit and supervisory goodwill.¹¹

C. Mitigation

A party injured by a breach of contract is expected to take reasonable steps to mitigate its loss. Long Island Sav., 60 Fed. Cl. at 89 (citing Wüliston § 64:27, at 191-200). Such a party may not recover “for loss that the injured party could have avoided without undue risk, burden or humiliation,” but “is not precluded from recovery ... to the extent that he has made reasonable but unsuccessful efforts to avoid loss.” Restatement (Second) Contracts § 350. ‘When mitigating damages from a breach, a party ‘must only make those efforts that are fair and reasonable under the circumstances.’ ” Home Sav., 399 F.3d at 1353 (quoting Robinson v. United States, 305 F.3d 1330, 1333 (Fed.Cir.2002)). Importantly, for mitigation the burdens of proof and persuasion shift from the plaintiff to the defendant. The burden is on the breaching party to show that the injured party’s efforts at mitigation were unreasonable. See, e.g., Franconia Assocs. v. United States, 61 Fed.Cl. 718, 741 (2004) (citing T.C. Bateson Constr. Co. v. United States, 162 Ct.Cl. 145, 319 F.2d 135, 160 (1963)).

Perhaps in an effort to avoid the consequences of this shift in the burden of proof, the government frames its arguments regarding mitigation in terms of causation and foreseeability, elements of the damage calculus for which the injured party bears the burden of proof. For example, the government argues that Globe’s losses were caused not by the government’s breach, but by Globe’s choice of an unreasonable method of mitigation. See Def.’s Post-Trial Reply Br. (“Def.’s Reply”) at 14 (“plaintiffs failure to raise capital ... is the immediate cause of any lost profits”).¹² This strategy has been rejected in prior decisions, most recently in Judge Allegra’s well-reasoned opinion in Franconia Associates, 61 Fed.Cl. at 749-51 & n. 60 (discussing Long Island Sav., 60 Fed.Cl. at 93). The court in Franconia correctly observed that where “the question presented involves an injured party’s failure to act to lessen damages, the issue does not involve proximate causation, but rather mitigation.” Id. at 750. Were the rule otherwise, the breaching party “would benefit from a pervasive causation defense not subject, like mitigation, to the bounds of reasoned decisionmaking. A breaching party simply could spin out a tangled web of damage-reducing hypothetical events in hopes that the injured party would become ensnared in failing to disprove at least one of them.” Id. Here, the government’s mitigation defense hangs on a very thin thread.

At trial the government called as an expert witness David J. Ross to testify about whether Globe’s decision to liquidate and not to raise replacement capital was reasonable. Tr. 4332:25 to 4333:5, 4355:19 to 4356:4 (Test, of Ross).¹³ According to Mr. Ross’s calculations, the cost to Globe of replacing its supervisory goodwill and capital credit by issuing debt securities would be $740,000, and the replacement of the same by issuing preferred stock would be $1.2 million. Tr. 4394:6 to 4397:5 (Test, of Ross); DX 1048. Mr. Ross reasoned that because Globe expected to realize profits in excess of the range of that cost, its decision not to raise the capital was unreasonable. Tr. 4406:12 to 4407:9 (Test, of Ross). Mr. Ross’s analysis is flawed as matters of economics, fact, and law.

Describing Mr. Ross’s premise, the government asserts that “the net cost of raising capital is limited to transaction costs.” Def.’s Br. at 24. That theory was squarely rejected by the Federal Circuit in LaSalle Taiman, 317 F.3d at 1374-75. Here, as in LaSalle, “[t]he government argues that the [replacement capital] had ‘no cost’ because ‘the firm receives cash worth [the same amount as the securities the firm issues],’ which is ‘the full value of the promise it is selling.’” Id. at 1374; see also Tr. 4395:21 to 4396:7 (Test, of Ross). The Federal Circuit concluded that the trial court “correctly rejected that argument, for capital is not ‘costless’ to either the investor or the recipient”; rather, “ ‘the cost of capital is the required rate of return on various types of financing.’” LaSalle Tai-man, 317 F.3d at 1374-75 (quoting James Van Horne & John M. Wachowicz, Jr., Fundamentals of Financial Management 387 (10th ed.1998)). In this respect, “the cost of capital does not depend on whether payment is made as debt, or out of anticipated future earnings.” Id. at 1375. See also Home Sav., 399 F.3d at 1352-1355 (affirming a trial court’s award of expectancy damages measured by cost of replacing capital, based on a thrift’s required rate of return for its capital financing less the “safe rate” of return the thrift could earn by investing the capital thus obtained).

Mr. Ross’s analysis does not supply evidence sufficient to sustain the government’s burden of proof on mitigation. The government relies on his testimony to contend generally that “to the extent a thrift believed it had profitable opportunities, and could convince investors and raise capital to pursue them, it could mitigate lost profits.” Def.’s Reply at 15. This hypothesis rests on a theoretical foundation divorced from the facts in this case. It would apply a severely restricted hindsight to the situation Globe faced in 1989 and 1990 and the actions it took. Instead, in the context of mitigation, ‘“reasonable conduct is to be determined from all the facts and circumstances of each case, and must be judged in the light of one viewing the situation at the time the problem was presented.’” Franconia Assocs., 61 Fed.Cl. at 741 (quoting In re Kellett Aircraft Carp., 186 F.2d 197, 198 (3d Cir.1950)). In the instant case, following FIRREA, the FDIC rated Globe a “5,” the worst possible rating, “reserved for institutions with an extremely high immediate or near term probability of failure.” Supra, at 341. Soon thereafter, the FDIC contacted Globe to arrange a liquidation examination. See supra, at 341-42. At the same time, OTS required Globe to submit a capital plan demonstrating how it would return to capital compliance. See supra, at 344. Tellingly, all of the four breach scenarios prescribed by Mr. Pittman and Ms. Fischman at FHLB-Topeka involved various degrees and ways of shrinking and two also assumed sale of branches; none assumed that Globe would attempt to raise capital. See supra, at 344. There is no evidence whatsoever in the record that any regulator urged Globe to raise capital rather than shrink. In all of the optional scenarios, Globe would not have been a profitable banking institution. See DX 122; see also supra, at 344, 347. Raising additional capital in those circumstances was not a plausible alternative to liquidation.¹⁴ In fact, the regulators recognized that their breach of the contract “caused management to downsize the institution,” PX 508 at 1, quoted supra, at 344, 346, and they praised Globe’s efforts to shrink drastically in an orderly way, opining that Globe had “sold off assets and deposit liabilities in impressive fashion to meet regulatory capital requirements.” PX 454 at WOQ2641376, quoted supra, at 345.

In short, the government has not discharged its burden to show that Globe’s actual efforts to mitigate its losses as a result of the breach by shrinking to the point of liquidation were anything other than reasonable under the circumstances.

D. Lost Profits

Indisputably, Globe was profitable at the time of the government’s breach. By June 30, 1989, the regulators reported that “[tjotal assets on a consolidated basis are approximately $735 million, which means that by rapidly increasing earning assets (and through FSLIC assistance), Globe has been able to achieve profitability.” PX 296 at 18 (Regulatory Plan (Sept. 22, 1989)). In its October 1989 report of examination, the FDIC noted that “[tjhrough the utilization of various hedging techniques, management has achieved a positive and stable net interest margin while effectively minimizing the institution’s overall interest rate risk. Year-to-date earnings are ahead of projections ($2,400,000 per year) and the loan loss reserve does not appear inadequate.” PX 306 at 1-a. Globe’s net interest income for fiscal year 1989 was $6.1 million, and, following the breach, was just below $6.0 million for fiscal year 1990. PX 397 at 4 (Letter from Deloitte & Touche to Globe’s Board of Directors (Aug. 10, 1990)) (attaching Consolidated Financial Statements and Supplemental Consolidating Schedules for the years ended June 30, 1990 and 1989 and Independent Auditors’ Report). Globe had received the regulators’ approval to grow to approximately $960 million during its initial three years of operation, and it was well en route toward such growth prior to the breach. See swpra, at 337.

On the facts proved at trial, this is a quintessential case for awarding expectancy damages in the form of lost profits. Damages measured as lost profits are the ordinary starting point for evaluating available remedies for breach of contract. As the Federal Circuit noted in Glendale Federal Bank, “[o]ne way the law makes the non-breaching party whole is to give him the benefits he expected to receive had the breach not occurred.” 239 F.3d at 1380 (citing Restatement (Second) Contracts § 344(a)). And further, “[t]he benefits that were expected from the contract, ‘expectancy damages,’ are often equated with lost profits, although they can include other damage elements as well.” Id. (citing Restatement (Second) Contracts § 347). See also Energy Capital, 302 F.3d at 1324-25; California Fed. Bank, F.S.B. v. United States, 245 F.3d 1342, 1349 (Fed.Cir.2001). Nonetheless, in the Winstar context, plaintiffs have had difficulty establishing a basis for an award of lost profits as expectancy damages. Of the three requisite elements for such damages, the first two facets, causation and foreseeability, have not been the stumbling block. Rather, the chief difficulty has proven to be the third element; i.e., that “a sufficient basis exists for estimating the amount of lost profits with reasonable certainty.” Energy Capital, 302 F.3d at 1325. Dicta in Glendale Federal Bank, FSB v. United States, 378 F.3d 1308 (Fed.Cir.2004), stated the problem in pithy, blunt terms:

Expectancy damages theory, based on lost profits, has proven itself impractical for these [Winstar ] cases, and generally not susceptible to reasonable proof. We have not, however, barred as a matter of law the use of expectancy/lost profits theory, but, given the speculative nature of such a damages claim, one that has yet to be successfully established in any Winstar case, experience suggests that it is largely a waste of time and effort to attempt to prove such damages.

Id. at 1313 (citations omitted). Yet, outside the Winstar arena, awards of lost profits have not been so disfavored as a remedy. Energy Capital was an action for breach of a contract with the Department of Housing and Urban Development (“HUD”) in which the Federal Circuit affirmed the trial court’s finding that the plaintiff had established its entitlement to lost profits. See Energy Capital, 302 F.3d at 1324-30. In particular, the court of appeals held that evidence at trial showed Energy Capital’s anticipated profits flowed directly from its agreement with HUD and did not stem “from ‘other independent and collateral undertakings.’” Id. at 1328 (quoting Wells Fargo Bank N.A. v. United States, 88 F.3d 1012, 1022-23 (Fed. Cir.1996)). Moreover, the court of appeals found the trial court’s computation of Energy Capital’s profits to be reasonably ascertainable and not speculative despite the fact that Energy Capital was a relatively new venture. Id., 302 F.3d at 1324-28.

More recently, even in the Winstar eases, awards of lost-profits damages have been made by trial courts and affirmed by the Federal Circuit when the facts adduced at trial supported such a result. See, e.g., LaSalle Talman Bank, F.S.B. v. United States, 64 Fed.Cl. 90, 100-106 (2005), on remand from 317 F.3d 1363 (Fed.Cir.2003) (thrift entitled to, among other things, $1.38 million for profits lost on shrink of asset portfolio necessary to achieve compliance with capital requirements); Commercial Fed. Bank, F.S.B. v. United States, 59 Fed.Cl. 338 (2004) (thrift entitled to damages of $5.602 million for profits lost due to shrinkage of assets to comply with FIRREA’s capital requirements), aff'd, 125 Fed.Appx. 1013 (Fed.Cir. 2005) (affirmance under Rule 36 of the Rules of the Court of Appeals for the Federal Circuit). In this vein, the analysis which follows focuses on whether, and the extent to which, lost-profits damages have been proved to a reasonable certainty in this case.

At trial, Globe put forward two sets of lost-profits damage calculations. First, W. Douglas Williams, a certified public accountant and the former chief executive officer of Globe, calculated both the profits Globe would have made during the 1990s absent the breach by continuing to operate according to its business plan and post-1999 lost profits determined on the basis of Globe’s residual branch value as of the end of 1999. For his calculation of profits lost during the 1990s, Mr. Williams relied on the assumptions that Globe would continue to follow its approved business plan and that it would continue to realize the operating results that Globe actually had achieved prior to the breach. Arithmetically, Mr. Williams incorporated these assumptions into a model that drew heavily upon a modified version of the projections made in April 6, 1990 as mandated by OTS-Topeka. Tr. 813:5-23 (Test, of Williams); PX 584 at 76-82 (Expert Report of W. Douglas Williams (Nov. 5,1999)). With respect to the OTS-mandated projections, Mr. Williams modified the scenario used for Exhibit A, the non-breach scenario, to incorporate more conservative assumptions by holding assets constant as of June 1992. Tr. 851:17 to 855:11, 858:2 to 873:10 (Test, of Williams); PX 584 at 80-82; PX 584-25; PX 584-20. Using these parameters, Mr. Williams calculated Globe’s profits lost during the 1990s to be $55.8 million. Tr. 893:5-10 (Test, of Williams); PX 584 at 82-83.

A more sophisticated set of calculations was prepared by a second expert witness, Andrew S. Davidson, whose firm advises financial institutions primarily in the area of mortgage-backed securities. Tr. 2259:2-15 (Test, of Davidson); PX 583 at 1 (Expert Report of Andrew S. Davidson (Nov. 5, 1999)). Mr. Davidson’s analysis addressed lost profits during the 1990s and was based on two principal components. Mr. Davidson first examined the extensive portfolio of mortgage-backed securities that Globe actually held at the date of the breach. Starting with the outstanding principal balances of the securities in that portfolio as of December 31, 1989, Mr. Davidson compiled the maturity “run-off” of each security on a month-by-month basis through August 31, 1999. Tr. 2325:15-20 (Test, of Davidson); PX 583 at 16-17, 20-21; PX 583-07. He used the actual outstanding balances whenever possible, but for the 20-30% of the portfolio for which such data were unavailable, he either used a proxy or made certain assumptions. Tr. 2325:15 to 2329:17, 2330:15 to 2331:20 (Test, of Davidson). Mr. Davidson then calculated the net-interest income that each security would have generated from January 1990 through August 1999 by multiplying the outstanding principal balance of the portfolio each month by the 98 basis point spread identified in the Smith Breeden Investment Committee Report for December 29, 1989. Tr. 2323:11 to 2324:8, 2325:4-14, 2335:8 to 2339:11 (Test, of Davidson); PX 583 at 16-17, 20-21; PX 583-07 (setting out the monthly net interest income in tabular form); PX 344, Ex. N. This analysis resulted in a total amount of $22,984,560 of net interest income generated over the 1990s from the existing portfolio of assets and duration-matched liabilities. Tr. 2324:9 to 2325:3, 2347:21 to 2348:8, 2351:3-9 (Test, of Davidson); PX 583 at 21; PX 583-07; PX 583-08.¹⁵

As a second step, Mr. Davidson developed a “reinvestment portfolio,” i.e., a portfolio of mortgage-backed securities that Globe could have purchased with the runoff from its actual portfolio. Because the government’s breach forestalled Globe from reinvesting the principal repayments, Mr. Davidson developed proxies for both the assets Globe would have purchased and the liabilities Globe would have used to fund those purchases. The net spread between those two proxies, the Lehman Mortgage Market Index and the cost of duration-adjusted swap agreements, respectively, generated a net interest margin for the reinvestment portfolio. Tr. 2365:4-9 (Test, of Davidson); PX 583 at 17, 22-24.¹⁶ With respect to the asset proxy, Mr. Davidson applied an upward adjustment of 25 basis points to account for Globe’s ability to select among more attractive securities than those included in the Lehman index and to choose the timing of its investments. Tr. 2370:18 to 2372:18 (Test, of Davidson); PX 583 at 24-26. More specifically, whereas the index is limited to “agency” securities, ie., securities issued by Ginnie Mae, Fannie Mae, or Freddie Mac, Globe could and did invest in adjustable-rate securities, insured non-agency mortgage-backed securities, and bonded tax-advantaged residuals. Tr. 2366:4 to 2368:15, 2372:25 to 2384:11 (Test, of Davidson); PX 583 at 25-26; PX 583-09. On the liability side, Mr. Davidson made adjustments to the swap market proxy to reflect Globe’s use of and continued access to lower-cost funds relative to LIBOR, such as deposits and FHLB-advances. Tr. 2399:19 to 2402:17 (Test, of Davidson); PX 583 at 26; PX 583-10. He applied a downward 25 basis-point adjustment (ie., one favorable to Globe) assuming a portfolio size of $700 million and a downward 20 basis-point adjustment assuming a portfolio of $1 billion. Tr. 2399:2-6 (Test, of Davidson); PX 583 at 26-27.

Using this approach, Mr. Davidson calculated that the net spread over the ten years would range from 60 to 115 basis points on a monthly basis, Tr. 2421:15 to 2424:4 (Test, of Davidson), with an average of 91 basis points on an annual basis. PX 583 at 29; PX 583-14.¹⁷ He determined that the reinvestment portfolio would have generated net interest income of $38 million on an asset base of $700 million and net interest income of $55 million if Globe had grown to $1 billion in assets. Tr. 2353:24 to 2355:24 (Test, of Davidson); PX 583 at 28. Combining the reinvestment portfolio and the actual portfolio, Mr. Davidson’s analysis resulted in net interest incomes of $61 million and $78 million on portfolios of $700 million and $1 billion, respectively. Tr. 2353:5-23, 2419:16 to 2421:8 (Test, of Davidson); PX 583 at 29-30; PX 583-12; PX 583-13.

1. Lost profits from 1990 to 1999.

The salient question is whether there is a sufficient basis in the evidence presented at trial to compute the amount of profits Globe lost from 1990 through 1999. To calculate Globe’s expenses of operations over the ten-year period, Mr. Williams relies on a modified form of the projections submitted to OTS in April 1990. He brings forward Globe’s expenses at year-end 1989 with an adjustment for inflation for each year. See PX 584-25. According to Mr. Williams’s analysis, Globe’s net tangible branch and operating expenses from January 1, 1990 through August 31, 1999 amount to $44,388,000. See id. The government does not strongly contest the expense side of his analysis, but rather it focuses its criticism on the income side of Mr. Williams’s projections. See generally Def.’s Br. at 27-50; Def.’s Reply at 1-11. The court has examined the expenses calculated by Mr. Williams and finds them to be appropriate to serve as a component of a damages calculation.

In contrast, the court accepts much of the government’s critique of Mr. Williams’s income projections. In particular, Mr. Williams in large measure took as a given the income-generating capacity of Globe’s portfolio as it existed in December 1989, and he extrapolated that capacity to a larger portfolio, as authorized by Globe’s approved business plan. He also extended that projection with very few adjustments to reach from January 1, 1990 through December 31, 1999. The court finds that the starting point for Mr. Williams’s analysis of the income side of his projections was viable but his extension of it to a larger portfolio over an ensuing ten-year period was not.

Mr. Davidson presented an alternative evaluation of Globe’s income-generating capacity by also starting with the actual portfolio Globe held at the time of the breach. As described previously, Mr. Davidson referred to publicly available data regarding the securities in the portfolio to determine the actual performance of that portfolio over the nearly ten-year period ending August 1999. See supra, at 352. He also constructed a reinvestment portfolio of mortgage-backed securities and liabilities of the type Globe used to generate income values for the runoff from the actual portfolio, that was slightly smaller ($700 million) than the $735 million portfolio Globe had in June 1989, and a reinvestment portfolio that was somewhat larger ($1 billion). See supra, at 353.

The government’s criticisms of Mr. Davidson’s analysis are four-fold. First, according to the government, Mr. Davidson’s analysis does not actually put forward a detailed calculation of lost-profits damages. See Def.’s Br. at 51-52; Def.’s Reply at 12. Second, the spreads exhibited by the existing portfolio would not remain valid for longer than one year. See Def.’s Br. at 52-56; Def.’s Reply at 12. Third, the analysis is speculative because prepayment risks are not adequately taken into account. See Def.’s Br. at 52-56; Def.’s Reply at 12. And, fourth, available spreads in the mortgage-backed market were narrowing as the market became larger and more competitive during the 1990s. See Def.’s Br. at 56-58. Each of the government’s arguments about Mr. Davidson’s analysis will be examined in turn.

First, the government’s initial criticism is unavailing because, although Mr. Davidson’s analysis does not by itself provide a basis for determining lost profits, it can serve to supply the income-based side of a calculation that could draw upon Mr. Williams’s expense figures to generate a value for lost profits.

Second, Mr. Musika testified on behalf of the government that Mr. Davidson’s analysis inappropriately applied a 98-basis-point spread taken from the Smith Breeden Investment Committee Report for December 1989 only to the asset side and ignored the liability side, and that this spread was not intended to be used beyond twelve months. Tr. 4773:7-14, 4781:15 to 4788:18 (Test., of Musika). In addition, Mr. Musika opined at trial that Mr. Davidson’s application of the 98-basis-point spread is unreliable because it fails to adjust for interest-rate movements. Tr. 4789:6 to 4790:25 (Test, of Musika). Mr. Musika’s criticisms in this regard are not accepted by the court.

Mr. Musika is correct that Smith Bree-den’s projection that Globe’s portfolio would generate a spread of 98 basis points in 1990 assumed that interest rates would remain constant. However, Smith Breeden’s interest rate shock analysis demonstrates that the portfolio would have generated a higher spread if interest rates declined, as they actually did between 1990 and 1993. See Tr. 2335:8 to 2347:13 (Test, of Davidson); PX 344, Exs. M & N. In particular, while Smith Breeden projected net interest income of $6.3 million assuming a stable rate, that amount would have increased up to $18.8 million in a falling interest rate scenario. Tr. 2344:7 to 2345:25 (Test, of Davidson); PX 344, Ex. M.

In light of the prepayment data regarding Globe’s portfolio, the court considered one type of lost-profits calculation not put forward by either of the parties, i.e., primarily relying on Globe’s existing portfolio as a basis for lost profits through July 1995, as contrasted to August 1999. Use of such an earlier cutoff would have reduced reliance on the reinvestment portfolio as a source of lost profits. However, because the court has recast the parameters governing the reinvestment portfolio in a manner that is extraordinarily cautious, and favorable to the government, use of the earlier cutoff would engender only a relatively moderate change in the lost profits calculated for the period from 1990 to 1999.¹⁸

Most importantly, it bears emphasis that Globe’s investment strategy did not depend on predicting interest rates but rather managing its net interest margin on a duration-based approach that protected Globe against wide swings in interest rates. Tr. 2040:9 to 2041:4 (Test, of Williams).

Third, notwithstanding the government’s prepayment argument, Mr. Davidson’s calculations showing the remaining principal balance on Globe’s portfolio over time indicates that prepayments were steady until approximately June 1992, when they accelerated, before slowing again in February 1994. See PX 583-06. The falling interest rates and consequent faster prepayment schedules from 1992 through 1994 would result in most of the $22,984,560 of net interest income being earned in the first five years, meaning that Mr. Davidson’s analysis does not realistically depend on application of the 98-basis-point spread over a ten-year period. Tr. 2342:7-17, 2347:21 to 2349:16 (Test, of Davidson); PX 583-08. The court finds that these circumstances are sufficient to support Mr. Davidson’s calculation that $22,984,560 represents a reasonably certain estimation of Globe’s profits lost on its existing portfolio.

Fourth, the more difficult issue is determining a reasonably certain amount of profitability from reinvestment. The evidence presented at trial demonstrates that the practical size of Globe’s portfolio under the final revisions to its approved business plan was about $900 million. Globe did not and could not stay fully invested in mortgage-backed securities up to its approved size of approximately $960 million for two reasons. First, Globe’s existing portfolio was continuously running off throughout time. See supra, at 352. Second, Globe operated pursuant to a policy of making relatively few investments and making such matched asset-liability investments only when its spread targets were met. Tr. 194:6 to 195:22, 367:24 to 376:23 (Test, of Williams); PX 42 at 18-19, Ex. 3 at 36-37 (Globe Business Plan (Oct. 24, 1986)). Because of this policy, Globe’s actual level of asset purchases fell below that projected in its business plan. For example, beginning with Globe’s Quarterly Business Plan Review for the quarter ending September 30, 1988, Globe informed the regulators that its “[ijnvestment in service corporations was approximately $14 million less than projected as a result of limited opportunities to purchase assets with acceptable spreads.” DX 207. Similarly, for the quarter ending December 31, 1988, Globe disclosed that such investments were “approximately $22 million less than projected as a result of continuing limited opportunities to purchase assets with acceptable spreads.” DX 208. For the same reason, Globe reported that its investments in the quarter ending March 31, 1989 “continued to be significantly less than projected.” DX 209. And in its Regulatory Plan Supervisory Profile as of September 22, 1989, OTS-Topeka reported that

[i]n recent months, the spreads available on investment assets that meet Globe’s criteria have declined significantly (this trend is due both to the flattening of the yield curve and the narrowing of many fixed income investment spreads to Treasury instruments). When opportunities are available in the current environment, they ordinarily entail net hedge spreads varying between 30 and 70 basis points, and in rare cases over 100 basis points. For most institutions, these spreads would not be sufficient to cover operating expenses on an overall basis.

PX 296 at 21. OTS-Topeka further elaborated that “[t]he reason that operating results [in terms of net operating income] have exceeded projections is because the spread on the current securities investments, although low, is higher than projected. This issue of supervisory concern is that, since spreads have narrowed, Globe is no longer able to attain similar spreads.” Id. at 22. The court finds that Globe could have succeeded with lesser spreads on its reinvestment portfolio, but its track record of caution in making investments would have led it to slow the pace of investment even further, to the point that it is not reasonably certain that Globe would have grown on an average basis to $900 million, let alone to its maximum permitted size of $960 million. Based on the evidence, the court determines that a more appropriate average size for Globe’s overall investment portfolio would have been $835 million, only $100 million over the maximum size it achieved of $735 million in its portfolio in June 1989 before the advent of FIRREA slowed its investment activity almost to a standstill. At times, in a favorable investment climate, Globe might have been able to approach a total portfolio that exceeded $900 million, but the evidence shows that it could not have consistently maintained that size and still have followed its disciplined adherence to its detailed business plan.

Turning to Mr. Davidson’s analysis of the spread available on the asset side of the reinvestment portfolio, his use of the Lehman Mortgage Market Index and application of an upward adjustment of 25 basis points are reasonable in the circumstances. At trial, Mr. Musika criticized Mr. Davidson’s upward adjustment to the Lehman index because, according to Mr. Musika’s analysis, Globe actually performed slightly below the index from the period November 1987 to June 1990. Tr. 4796:23 to 4798:6, 4799:3-19, 4801:11 to 4804:18 (Test, of Musika); DX491, Ex. 8; DX 1081. To a similar effect, another government expert, Professor Fabozzi, testified at trial that the spreads on securities like CMO floaters that were attainable in the 1980s were generally not available in the 1990s, because more sophisticated players entered the market and increased efficiencies. Tr. 3500:13 to 3509:10 (Test, of Faboz-zi). See also Tr. 2811:22 to 2816:1 (Test, of Curry). However, neither Mr. Musika nor Professor Fabozzi took into account Globe’s ability to acquire up to approximately $275 million of bonded tax-advantaged residuals, which Globe did not begin purchasing until February 1989. Tr. 361:4-9, 366:10 to 367:23 (Test, of Williams). As reflected in Smith Breeden’s profitability study for December 1989, Globe’s CMO tax-advantaged residuals were yielding 180 basis points more than the total interest income on all of Globe’s rate-sensitive assets. Tr. 725:25 to 726:25 (Test, of Williams); PX 344, Ex. N; see also Tr. 360:19 to 361:3, 408:9 to 411:4 (Test, of Williams); Tr. 2378:20 to 2382:10 (Test, of Davidson). Furthermore, because these instruments had the effect of refreshing Globe’s NOLs, Globe would have been able to continue purchasing tax-advantaged residuals throughout the 1990s. Tr. 401:11 to 406:23 (Test, of Williams); Tr. 2382:11 to 2384:11 (Test, of Davidson).

Mr. Davidson testified that whereas Globe’s ability to purchase bonded tax-advantaged residuals would support an upward adjustment to the Lehman index, his adjustment of 25 basis points was based on options available to Globe under its business plan that did not include the prospect of increased investments in bonded tax-advantaged residuals. Tr. 2377:16 to 2378:19. Based on all of the evidence presented at trial, the court finds that a 25-basis-point adjustment over the Lehman index is appropriate based in large part on Globe’s demonstrated successful strategy of relying upon bonded tax-advantaged residuals as a significant portion of its portfolio.

Respecting the liability side of Mr. Davidson’s analysis, however, the court has determined that an adjustment of 25 basis points downward from LIBOR (favorable to Globe) is not warranted by the record in this case. Mr. Davidson based his analysis on an evaluation of the cost of funds available to Globe on the liability side. For a portfolio of $700 million, he assumed deposits of $200 million, FHLB advances of $350 million, and reverse repurchase agreements of $150 million. PX 583 at 27. For a portfolio of $1 billion, he assumed deposits of $200 million, FHLB ad-vanees of $500 million, and $300 million in reverse repurchase agreements. Id. Mr. Davidson had calculated Globe’s actual costs of funds for these liabilities, and he had concluded that Globe’s average cost of deposits was 62 basis points less than LIBOR between July 1989 and June 1990. Id. FHLB advances were available at 15 to 20 basis points below LIBOR, and “reverse repurchase agreements [were] generally priced right at or very close to LIBOR.” Id. Based on an average overall portfolio size of $835 million, however, the court has concluded that Mr. Davidson probably was slightly optimistic in his assessment both of the composition of the liability side of a Globe portfolio and the rate charged for FHLB advances.

The one relatively fixed source of funds for Globe was its deposit base. That base approximated $200 million in size, as Mr. Davidson assumed. And, there is no basis to reject Mr. Davidson’s calculation that Globe’s cost of deposits averaged 62 basis points less than LIBOR. For the sake of prudence, the court will round this cost advantage downward to 60 basis points. Similarly, there is no basis to reject reverse repurchase agreements as an available source of funds, at LIBOR. Globe had carefully entered into agreements with a number of investment banking houses to secure that availability, and it had demonstrated its ready reliance on such funds in building its portfolio. See supra, at 337. However, those funds were available at LIBOR, not below it. Lastly, FHLB-Topeka had proved to be a reliable source of funds for Globe, and the court credits the evidence that such funds were available below LIBOR, but not as far below as Mr. Davidson projected. Rather than considering FHLB advances to have been available 15 or 20 basis points below LIBOR, it seems prudent to consider them to have been available at 10 basis points below LI-BOR. Finally, it seems appropriate to put a somewhat heavier emphasis on revenue repurchase agreements than did Mr. Davidson, primarily because of the existence of Globe’s agreements with investment banking houses and the resulting somewhat greater assur-anee that reverse repurchase agreements as a source of funds would be available at LI-BOR rates as contrasted to the somewhat lesser likelihood that FHLB advances would always be available at 10 basis points less than LIBOR.¹⁹

Constructing a liability portfolio on this basis with $200 million derived from deposits, $180 million obtained through FHLB advances, and $455 million generated by reverse repurchase agreements results in a weighted average cost of funds of roughly 16.53 basis points below LIBOR. Rounding downward to 15 basis points to add a further measure of certainty produces a cost of funds to Globe of 15 basis points less than LIBOR.

In sum, the court concludes that the reinvestment portfolio derived by Mr. Davidson should be recalculated using an overall size of $835 million plus a spread of 40 basis points over the returns on the Lehman index. Globe shall provide this recalculation by May 27, 2005, and the government shall respond to Globe’s proffered calculation by June 13, 2005. Thereafter, the court will enter a judgment in this case that includes an award of lost profits for the period from January 1, 1990 through August 31, 1999, that adds the $22,984,560 derived from Globe’s existing portfolio and the amount to be calculated from the reinvestment portfolio, and then subtracts $44.388 million as Globe’s net tangible branch and operating expenses. The court finds and concludes that such an award satisfies the reasonable certainty test for awarding lost profits in this Winstar case, given the unique facts and circumstances associated with Globe’s inception, proven experience in operation under its business plan, and very substantial existing portfolio as of the date of the breach.

2. Postr-1999 lost profits.

As the measure of lost profits for the period after 1999, Mr. Williams considered the value that Globe’s branch network would have had as of 1999. His premise for that assessment was that Globe would have followed its fundamental business plan to shift to operations as a traditional thrift after the Oklahoma economy had recovered and Globe had achieved a satisfactory tangible-capital position. Tr. 178:24 to 179:4,198:20 to 199:3, 340:10 to 341:2 (Test, of Williams). Mr. Williams based his calculations on Globe’s deposit base as of 1990, the date of its sale, projected forward to 1999, plus market data for deposit premiums of contemporaneous branch sales in Oklahoma. Relying on the FDIC Branch Office Data Book, Mr. Williams found that the deposits in the six branches that Globe had sold to MidFirst totaled $189.2 million as of June 30, 1998. Tr. 901:2 to 902:16 (Test, of Williams); PX 584r-29. He then added to that number $8.09 million of deposits in the Harrah branch that Globe had finally closed in 1993, resulting in a total of $197.3 million of deposits that Globe would have held in 1999 but for the breach. PX 584 at 85; PX 584-29.

For purposes of calculating a reasonably ascertainable deposit premium, Mr. Williams examined branch deposit sales that were completed in Oklahoma between January 1998 and October 1999 and for which deposit premium data were available. Tr. 910:17 to 914:5 (Test, of Williams); PX 584 at 85; PX 584-30. He calculated the average deposit premium to be 6.75 percent. Tr. 910:17 to 914:7 (Test, of Williams); PX 584 at 85; PX 584-30. Applying that ratio to $197.3 million of total deposits in the seven branches, Mr. Williams determined that Globe’s residual value in 1999 would have been $13,318,290. Tr. 915:3-16 (Test, of Williams); PX 584 at 85; PX 584-31.

In response to Mr. Williams’s calculation, the government argues that “it does not even approach proving that [Globe and MidFirst], as a whole, would have been equivalent throughout the 1990s, when the yardstick data were generated.” Def.’s Reply at 17. The government relies on Mr. Bankhead’s testimony at trial that Globe’s deposit base had been shrinking prior to its sale of the branches, MidFirst had “a much more multidimensional business strategy than that of Globe,” and “consequently [MidFirst] was better positioned to grow deposits than Globe was or than Globe would have been.” Tr. 4010:23 to 4011:2; see also Tr. 3998:5 to 4014:25 (Test, of Bankhead); DX 510 1HI 20-21 (Bankhead Supp. Report (Feb. 24, 2004)).

As a factual matter, Globe has the better position in this dispute. Deposits in the six branches did not change markedly during their operation by MidFirst. Tr. 901:2 to 903:1 (Test, of Williams). In MidFirst’s hands, deposits for the branches continued to decline from 1990 through 1994. See DX 506B. Given the state of the Oklahoma economy during this period, this result is readily explicable. As Mr. Bankhead testified, “[deposits in thrifts headquartered in the [S]tate of Oklahoma declined during the late ’80s and ’90s.” Tr. 4269:12-14. And, over the longer period of the decade, from 1990 to 1999, three of the branches declined absolutely (Shawnee I, Pauls Valley, and Stroud). See id. Only two of the six branches, the Yukon and Shawnee II branches, showed substantially higher deposits. See id.; PX 584-29. The other branch, El Reno, increased its deposits very modestly. See DX 506B.

Mr. Bankhead emphasized that Globe’s overall deposit base had declined 22.3 percent in the two years prior to its sale to MidFirst, and that its core deposits had declined 33.5 percent in that same period. DX 510 H 20. He criticized Globe’s operation of its branch network, opining that “Globe placed little emphasis on traditional thrift activities such as single-family mortgage lending and deposit gathering.” Id. Mr. Bankhead is correct insofar as the change in Globe’s deposits during 1987-89 is concerned, but he is wrong as to the cause of the downward shift. Globe had an active program to “run off’ the jumbo, brokered deposits that had been used by OK Federal and Shawnee in a last-ditch effort to supply funds for increasingly unprofitable, money-losing operations. Tr. 323:2 to 325:7, 334:8 to 335:3 (Test, of Williams); Tr. 3989:6-19, 4252:11-18 (Test, of Bankhead). In effect, Globe was shedding its expensive hot-money certificates of deposits and retreating to a more stable core of locally-generated savings deposits and smaller certificates of deposit. Tr. 4240:21-23 (“[S]ome of the out-of-market CDs and the jumbo CDs, the broker deposits were gone. That’s your more volatile money.”) (Test, of Bankhead). In addition, Mr. Bankhead’s definition of core deposits is much too narrow. He defined core deposits to be total deposits less escrow amounts and less certificates of deposits. Tr. 4077:14-20 (Test, of Bankhead). Mr. Bankhead acknowledged that definitions of core deposits other than his own included certificates of deposit under $100,000, or, alternatively, under $80,000. Tr. 4081:4-6 (Test, of Bank-head). He also indicated that “a core deposit is better defined as something that’s stable and will remain in the institution rather than by a dollar amount.” Tr. 4080:25 to 4081:2. In its deposit base, Globe in fact consistently had a large proportion of relatively small certificates of deposits, but these were excluded by Mr. Bankhead from core deposits as he defined the term. See DX 1041. Globe’s internal financial reports broke down the deposits by type, as did the thrift financial reports that it filed with the financial regulators. Tr. 4081:24 to 4082:5 (Test, of Bankhead). On July 31, 1998, Globe’s average certificate of deposit was approximately $10,000 in size, and that average amount for its certificates of deposit remained relatively constant throughout its operation. Tr. 4082:13 to 4083:5 (Test, of Bankhead). In short, Globe’s certificates of deposit were of small average size, stable, a significant proportion of total deposits, and as a low-cost funding source were important to Globe in its overall thrift operations.

Importantly, MidFirst’s operation of the branches did not differ significantly from that of Globe. MidFirst had a large mortgage-servicing operation and in the course of that business received significant payments into escrow accounts held for mortgage borrowers for the periodic payment of taxes and insurance. Tr. 3993:4 to 3995:11, 4244:9-20 (“MidFirst was unusual in the sense that they had a very stable balance of escrow deposits____MidFirst had service loans all over the country____[M]oney comes in and out of a transaction account, just like it does an escrow account. It builds up over time, and then when the taxes or insurance is paid, money will come out of that escrow account.”) (Test, of Bankhead). Mr. Bank-head resisted any suggestion that MidFirst was focused on non-traditional thrift operations, Tr. 4271:2-19, but he had to concede on cross-examination that less than ten percent of MidFirst’s assets were comprised of mortgage loans. Tr. 4271:20 to 4274:7 (Test, of Bankhead). In fact, in the early 1990s, most of MidFirst’s assets consisted of mortgage-backed securities just as Globe’s were. Id. Consumer loans were not a significant factor. They made up a small proportion of Mid-First’s assets in the early 1990s although that aspect of its operations grew larger as the decade progressed. Id. Like Globe, Mid-First used wholesale funding and was a Smith Breeden client. Tr. 4273:21 to 4274:7 (Test, of Bankhead). Apart from MidFirst’s mortgage-servicing activity, the substantial difference between MidFirst’s operations and those of Globe was that MidFirst did originate mortgage loans, which it then sold and securitized. Tr. 4273:17-20 (Test, of Bank-head).

In short, Globe’s branch network was still in place throughout the 1990s, operated by MidFirst under relatively comparable conditions, and the deposits in those branches did not change appreciably during their operation by MidFirst. Thus, the branch network Globe sold, taken as of 1999, provides a factually sound foundation for deriving the amount of Globe’s residual value with reasonable certainty.

As to Mr. Williams’s methodology for his calculation of a deposit premium, the government puts forward two arguments. First, Mr. Bankhead testified that “[sjimply taking the average [deposit premium] is not necessarily going to get you the number that’s going to be representative” because “[b]ranches and deposits just aren’t created equal.” Tr. 4022:7-10. There is some validity to this argument because in 1998 and 1999 there were relatively few branch sales in Oklahoma for which deposit premiums were reported — mainly those Mr. Williams addressed. See Tr. 4116:12-19 (Test, of Bank-head). At that time, the Oklahoma economy had improved, and there were not any “distress” sales of branches as there were during the period from 1989 to 1992. Tr. 4105:1 to 4113:23 (Test, of Bankhead) (sale of approximately 40 branches in that period, some of which were by the RTC). The limited sam-pie for 1998 and 1999 is sufficient for analytical purposes, representing as it does the actual market conditions for branches in Oklahoma at the relevant time. However, sales of branches that were small in size might unduly affect a straightforward arithmetic average of the deposit premiums paid for branches. To ensure that a deposit premium is derived that reflects the actual amount of deposits sold rather than numbers of branches, the branch sale premiums could be weighted by size of deposits in each branch sold, as follows:

Buyer Seller Deposits Premium/ Weighted ($000) deposits (%) summation

ANB Bankcorp Midland Financial Corp.²⁰ 13,400 7.10 951.4

Bank of Cushing & Trust Co. BaneFirst Corp. 9,523 7.00 666.61

BOK Financial Corp. BaneFirst Corp. 32,727 6.50 2,127.255

Falcon Bancorp, Inc. BaneFirst Corp. 15,500 5.81 900.55

Hall Properties BaneFirst Corp. 8,018 7.08 567.6744

NBC Bankshares Pawhuska Inc. BaneFirst Corp. 7,000 7.00 490

86,168 5703.4894

See PX 584-30. This weighted calculation produces a premium of 6.62 percent. Applying this ratio to the $197.3 million of total deposits produces a market-based value of $13,061,260 for the residual post-1999 ex-Globe branches.

Second, Mr. Bankhead observed that five of the six transactions on which Mr. Williams relied involved sales by a bank, not a thrift, and he opined that bank premiums tend to be higher than thrift premiums because, among other things, he said banks generally have a higher core deposit ratio. Tr. 4020:18-21, 4022:24 to 4023:19. In particular, Mr. Bank-head testified that BaneFirst, the bank that sold branches in those five transactions, had a core deposit ratio of almost 60 percent as of December 31, 1998, whereas Globe’s ratio was only 12 percent prior to its sale to MidFirst. Tr. 4023:20 to 4025:4. In addition, he testified that BaneFirst was able to grow its deposits fairly regularly. Tr. 4023:25 to 4024:22; DX 506, Ex. F.

The government’s position does not account for the salient fact that the highest of the six premiums was realized on the sale of one of Globe’s former branches, MidFirst’s Stroud branch, which manifestly was not a bank branch. As Mr. Bankhead recognized, Midland Financial Corporation, the seller in the first row in the table supra, was the holding company for MidFirst and sold Globe’s former Stroud branch in that transaction. Tr. 4020:24 to 4021:9 (Test, of Bank-head); see also Tr. 914:12-24 (Test, of Williams). In response to this evidence, the government again refers to Mr. Bankhead’s testimony about “the vast differences” between Globe and MidFirst, arguing that “it is speculative to assume Globe, in a non-breach scenario, could have obtained a similar deposit premium.” Def.’s Br. at 65 n. 25. This argument is again undercut by the fact that MidFirst obtained relatively poorer results than Globe during its actual operation of the Stroud branch. Midland’s (MidFirst’s) Stroud transaction was completed on July 29, 1998. PX 584-30. Between December 31, 1989 and June 30, 1998, deposits in Stroud had increased by only $637,000 despite the fact that the Oklahoma economy had improved substantially in the latter part of the decade. See DX 506B. Moreover, during MidFirst’s operation of the Stroud branch between 1998 and 1999, deposits declined by $1,039,000. See id.

The court finds that a deposit-weighted analysis of available data for the branch sales completed in Oklahoma between January 1998 and October 1999 provides an appropriate basis for determining a deposit premium. Most importantly, on this record it is reasonably certain that Globe would have generated' profits after 1999 but for the government’s breach and that the amount of $13,061,260 represents a reasonably certain approximation of those profits. In that regard, the court is guided by the Federal Circuit’s “remark that when damages are hard to estimate, the burden of imprecision does not fall on the innocent party.” LaSalle Talman Bank, 317 F.3d at 1374. See also Restatement (Second) Contracts § 352 cmt. a (“Doubts [about reasonable certainty] are generally resolved against the party in breach____Damages need not be calculable with mathematical accuracy and are often at best approximate----This is especially true for items such as loss of good will as to which great precision cannot be expected.”).

3. Offset for incidental losses.

As the foregoing sections indicate, this court will award lost profits in this case, measured by (1) computations based upon Globe’s actual portfolio and upon its runoff investment portfolio for the period from January 1, 1990 through August 31, 1999, and (2) computations of the residual value of Globe’s branch network as at August 31, 1999. However, as previously indicated, the parties separately identified incidental losses as a category of damages to be awarded under Restatement (Second) Contracts § 347, in addition to lost profits. The court has accordingly also undertaken a separate analysis of incidental losses, as set out in the next section. Nonetheless, because of the methodology used by the experts for each side, the analysis of lost profits also bears on incidental damages. In this regard, Globe concedes that “[t]he incidental damages overlap with the past lost profits. Thus, to the extent the [c]ourt awards incidental damages, [plaintiffs’ claims for lost profits would be reduced dollar-for-dollar.” Pis.’ Br. at 43. The court will implement this concession. It will award incidental damages separately from lost profits in the judgment but will offset the amount of that award in determining the award of damages for lost profits.

E. Incidental Losses

Globe seeks as incidental damages four categories of out-of-pocket costs it argues it would not have incurred but for the government’s breach. These costs are termed “incidental losses” by the Restatement (Second) Contracts, and they protect the injured party’s expectation interest but are separate and distinct from lost profits. Restatement (Second) Contracts § 347 cmt. e (“Items of loss other than loss in value of the other party’s performance are often characterized as incidental or consequential.”).

1. Prepayment penalties on FHLB advances.

One of the ways Globe chose to mitigate its loss in connection with its shrinkage to liquidation was to terminate prematurely some of its FHLB advances. In that connection, Globe paid prepayment penalties to FHLB-Topeka in the amount of $2,150,890. Tr. 792:19 to 807:5, 839:6 to 842:9 (Test, of Williams); DX 321 at PGBO 13 0314; PX 526 at 3, 15-16 (Phoenix’s consolidated financial statements for the years ended June 30,1992 and 1991 and independent auditors’ report).

The government does not dispute the numerical basis for Globe’s prepayment-penalty claim. Tr. 4634:14-18 (Musika testimony) (“The amounts that plaintiff came up with are indeed accurate amounts”). Instead, the government argues that these costs would have to be offset by the corresponding gains on liabilities and assets and that in any event, because Globe terminated its FHLB advances after achieving capital compliance, the breach could not have been the proximate cause of Globe’s payment of the associated penalties. Def.’s Br. at 66-68; Def.’s Reply at 18-19. Both contentions by the government are unavailing.

With respect to the first argument, the calculation of Globe’s lost profits set out above already takes into account both the losses incurred on the liability side and the gains in assets. However, Globe’s incidental damages will be excluded from the lost-profits calculation to avoid double-counting. Regarding the government’s contention that “[i]t is nonsensical to assert that FIRREA caused post-compliance shrinkage,” Def.’s Reply at 19, the court again rejects that argument for the reasons explained supra, at 347.

2. Early termination fees on interest rate swaps.

Globe also shrank by terminating prematurely a number of its interest rate swap agreements, incurring fees of $4,278,142. Tr. 775:9 to 792:12 (Test, of Williams); PX 397 at 4 (Letter from Deloitte & Touche to Globe’s Board of Directors (Aug. 10, 1990)) (attaching Consolidated Financial Statements and Supplemental Consolidating Schedules for the years ended June 30, 1990 and 1989 and Independent Auditors’ Report); PX 475 at 3, 16-17 (Deloitte & Touche consolidated financial statements for Globe for years ended June 30, 1991 and 1990 and independent auditors’ report). The government does not dispute this amount but rather has extended to this claim its criticism of Globe’s claims for prepayment penalties on FHLB advances. See Def.’s Br. at 66-68; Def.’s Reply at 18-19. Those arguments are equally unpersuasive here. See supra, at 347. Globe’s claim for early termination fees on interest rate swaps is allowed.

3. Deposit disposition costs.

Globe also seeks as incidental damages the difference between the amount FSLIC paid Globe to assume the OK Federal and Shawnee deposits on the one hand, and the amount Globe paid MidFirst to accept those deposits, taking into account Globe’s liquidation of the Harrah deposits, on the other. To calculate these damages, Globe added $6.8 million of goodwill to the $3.7 million premium on the Shawnee deposits for a total of $10.5 million. Pis.’ Br. at 39-40. Plaintiffs then divided that amount by $229 million of deposits acquired in the two transactions to determine an average premium of 4.58 percent. Id. at 40. Subtracting from that figure the 0.7 percent premium paid to Mid-First, they derived a premium of 3.85 percent. Id. at 41. Applying the 3.85 percent premium to the $165.4 million of deposits sold to MidFirst, plaintiffs computed $6.4 million of damages. Id. Next, because Globe paid a premium of 2.61% when it assumed the $8.09 million of Harrah deposits but redeemed them at par when it repaid those depositors, Globe incurred a cost of $211,000. Id. Adding the $6.4 million costs of disposing the OK Federal and Shawnee deposits to the $211,000 cost of liquidating the Harrah deposits results in a total claimed amount of approximately $6.6 million of incidental damages. Id.

The government responds that Globe improperly uses as a component of incidental damages the $6.8 million of goodwill Globe booked when it acquired OK Federal. Def.’s Br. at 69-70; Def.’s Reply at 20. This court agrees. It has rejected a similar attempt by plaintiffs in an earlier proceeding in this case to treat the loss of goodwill as a loss of paid-in cash capital. See Globe Sav., 59 Fed.Cl. at 98 (citing LaSalle Talman, 317 F.3d at 1376; California Fed. Bank, 245 F.3d at 1350-52; Glendale, 239 F.3d at 1381-83). Loss of goodwill is not an incidental cost, and goodwill should be eliminated from the calculation of deposit disposition costs. With this elimination, the calculation of Globe’s loss on disposition of its deposits is simplified. Subtracting the 0.7 premium gained on the Mid-First transaction from the 2.61 premium paid on the Shawnee transaction, results in a lost premium of 1.91 percent. Applying this premium loss of 1.91 percent to the $165.4 of deposits Globe divested to MidFirst produces an amount of loss of $3,159,140. Adding to that amount the $211,000 lost from the pay off of the Harrah deposits results in total incidental damages of $3,370,140 on Globe’s disposition of deposits.

The government correctly observes that awarding both the claim for deposit disposition costs and the claim for the residual value of the branches would constitute double-counting. Def.’s Reply at 20 n. 14. The court will ensure that the damages awarded to Globe in this case address the deposit disposition costs of $3,370,140 as incidental damages, not as lost profits.

4. Expenses associated with preparing capital plan.

In his expert report, Mr. Williams claimed as “other losses” Globe’s cumulative net loss of $482,000 for the period from 1990 through June 1993. PX 584 at 82-84; see also PX 584-28. At trial, plaintiffs’ counsel stipulated to reduce the amount of this claim to $67,000, in response to the government’s observation that $150,000 of the original amount represented legal fees and that $265,000 represented bonuses paid to Globe’s executives. Tr. 1527:9 to 1533:22, 1545:5 to 1546:19. The resulting claim of $67,000 in “other costs” includes expenses in connection with preparing and presenting to the regulators Globe’s capital-exception request, the financial projections under the five scenarios specified by FHLB-Topeka, and finally a capital plan that was approved by OTS-Topeka. See supra, at 343^44. Although there is no doubt that Globe incurred expenses in this regard, Globe’s proofs did not identify these expenses with particularity. Applying the standard of reasonable certainty to this claim for damages, the court finds that a two-thirds reduction from the claimed amount is appropriate. Accordingly, only $22,333 is allowed as incidental damages for these “other costs.”

F. Discounting for Present Value

Lastly, the government argues that all of Globe’s damages calculations are flawed because they fail to discount to the date of the breach of contract. Def.’s Br. at 31-34. In support of this contention, the government purports to rely on the Federal Circuit’s ruling regarding discounting damage awards for breach of contract in Energy Capital, 302 F.3d at 1330. See Def.’s Br. at 31-34. Energy Capital, however, supports Globe’s position, not that of the government. In that case the court of appeals observed that “the damages that would have arisen after the date of judgment (‘future lost profits’) must be discounted to the date of judgment.” 302 F.3d at 1330 (emphasis added). The court specifically rejected the government’s contention that the contract damages award should have been discounted to the date of breach. Id. (“[W]e hold that the trial court did not err in discounting Energy Capital’s lost profits to the date of judgment instead of the date of breach.”). See also Northern Helex Co. v. United States, 225 Ct.Cl. 194, 634 F.2d 557, 564 (1980) (discounting anticipated profits to the date of judgment). Here, all of the damages will antedate the date of judgment. Thus, there is no legal basis for discounting Globe’s damages in this ease.²¹

CONCLUSION

Based on the proofs at trial in this case, Globe is awarded expectancy damages for breach by the government of its contractual obligation to allow Globe to include capital credits and supervisory goodwill in its regulatory capital, to be amortized over a specified period of years. The regulators entered into a transaction with Globe and Phoenix to sell and transfer troubled thrift assets in Oklahoma, based upon a proposal by Globe that it be allowed to use the resulting intangible regulatory capital as a platform for a risk-controlled arbitrage program.. That proposal made sense for Globe and the government at the time, given the severely depressed Oklahoma economy, the lack of any other potentially interested purchaser of the assets, and the close, continuing controls and checks that the Bank Board imposed on Globe’s implementation of that business plan. All aspects of Globe’s implementation of its business plan were subject to approval by, and close scrutiny of, the regulators. Globe’s management strictly adhered to the approved business plan, was very selective in its asset acquisitions and duration-matched borrowings, and consistently and carefully limited risks in its implementation of the plan. Globe built a substantial, continuously profitable portfolio of mortgage-backed securities and duration-matched borrowings, which the government’s breach caused it to liquidate. Globe’s well defined portfolio provides the foundation for an award of lost profits as expectancy damages.

For the reasons stated in this opinion, the court awards Globe lost profits for the period from January 1, 1990 through August 31, 1999, based on (1) the portfolio built by Globe plus (2) an amount for reinvested runoff, less (3) the expenses of operating its thrift branch network. In this connection, the court has resolved the net interest income attributable to the existing portfolio and the expenses of sustaining the thrift branch network. The third component of this lost profits award, the results from reinvestment, remains to be determined. As to reinvestment, the court has specified the parameters to be used to calculate the net interest income attributable to the reinvestment portfolio. By May 27, 2005, Globe shall provide a calculation of the net income derived from the reinvestment portfolio using the parameters specified by the court. On or before June 13, 2005, the government shall submit any suggested corrections or alterations to the calculation provided by Globe.

As residual lost profits after August 31, 1999, the court awards $13,061,260 based on the projected value of Globe’s branch network as at that valuation date.

The court also awards Globe damages for its incidental losses in mitigating the breach, amounting to $9,821,505. As explained above, the methodology employed by Globe’s experts for calculating lost profits incorporated these incidental losses into the lost profit calculations. Therefore, the lost profits ultimately awarded shall be reduced by the amount of the incidental losses separately awarded as damages.²²

IT IS SO ORDERED.

1 . See United States v. Winstar Corp., 518 U.S. 839, 116 S.Ct. 2432, 135 L.Ed.2d 964 (1996).

2 . For these purposes, "supervisory goodwill” is defined as “the excess of the purchase price paid for a thrift over the fair value of all identifiable assets acquired.” Home Sav. of America v. United States, 399 F.3d 1341, 1345 n. 1 (Fed.Cir. 2005) (citing Winstar, 518 U.S. at 848-49, 116 S.Ct. 2432).

3 . The recitation of facts that follows constitutes the court’s findings of fact pursuant to Rule 52(a) of the Rules of the Court of Federal Claims ("RCFC”).

4 . The business-plan revision approved by FHLB-Topeka in March 1, 1988, just prior to the Shawnee transaction, had capped Globe's asset growth at $923 million. See DX 204 (Business Plan Supplement (Dec. 1, 1987)); PX 149 (Letter from R. Renz to W. Williams (Mar. 18, 1988)) (approving the business plan as revised and supplemented).

5 . These totals for the three liability categories represented Globe's liability composition in general terms as of June 1989. The combined amount shown for these categories is $694.4 million. Globe's balance sheet for the year ending June 30, 1989 reports total liabilities of $702.204 million. See PX 584-7. Globe’s tangible capital at the time was negative $1.439 million. Id.

6 . The financial regulatory entities employ a rating system, first known under the “MACRO" acronym and then, after FIRREA, as the “CAMEL” system. See Long Island Sav. Bank, F.S.B. v. United States, 60 Fed.Cl. 80, 86-87 n. 11 (2004); Southern Cal. Fed. Sav. & Loan Ass’n v. United States, 57 Fed.Cl. 598, 605-06 (2003). The MACRO assessment addressed the effectiveness of the management and the board of directors, and asset quality, capital adequacy, asset/liability and risk management, and earnings (operations). The CAMEL assessment referred to the same topics, but described them as capital, assets, management, earnings and liability. A rating of "1” was and is the strongest, reflecting that the institution is "sound in every respect ... and give[s] no cause for supervisory concern.” 61 Fed.Reg. 67021, 67025-26 (Dec. 19, 1996). A rating of "2” indicates that the institution is "fundamentally sound ..., [and] supervisory response is informal and limited." Id. at 67026. A rating of “5” is the weakest, indicating that the institution exhibits "extremely unsafe and unsound practices or conditions ..., [o]ngoing supervisory attention is necessary ..., [and] failure is highly probable.” Id.

7 . FIRREA mandated new capital standards as follows: “tangible” capital was to be maintained at a level "not less than 1.5 percent of the savings association's total assets," "core” capital was required to be "not less than 3 percent” of total assets, and "risk-based” capital was required to be kept at a level not "materially” lower than that required for national banks. 12 U.S.C. § 1464(t)(2). Supervisory goodwill and other unidentifiable intangible assets could not be counted towards tangible capital and were to be phased out of calculations for "core” capital by 1995. 12 U.S.C. § 1464(t)(3)(A), (t)(9)(A)-(C).

8 . Mr. Pittman’s directive reflected OTS’s guidance on interest rates that were to have been used in financial projections in a thrift’s capital plan. See Tr. 435:1-19, 1059:23 to 1064:6 (Test, of Williams), 3117:22 to 3119:11 (Test, of Pittman). In particular, Thrift Bulletin 36 ("TB 36") provided that such “projections should be based upon the continuation of the existing interest-rate and regional economic environments.” DX 449 at 2 (TB 36 (Nov. 6, 1989)). OTS provided further detail on this matter in Thrift Bulletin 36-1 ("TB 36-1”), which states, in pertinent part:

Presently, TB 36 requires financial projections to be based on existing interest rates. TB 36 does not specify the date of the “existing interest rates” on which financial projections should be based. To simplify and streamline the review process, the recommended "as of" date for interest rates used in the initial projections should be November 6, 1989. District Directors may approve capital plans from savings associations that have already prepared capital plans using interest rates as of a different date, but such plans should be based on interest rates that are not materially different from those that prevailed on November 6, 1989. Capital plans submitted after January 8, 1990, should be based on interest rates prevailing as of the last business day of the quarter ending prior to the date of submission.

DX450 at 1 (TB 36-1 (Dec. 14, 1989)). TB 36-1 also provided an implied interest-rate forecast through the fourth quarter of 1994. Id. at 1 & Table 1.

9 . Globe had previously considered the possibility of raising capital from existing shareholders or by merger with a stronger thrift or bank. See PX 296 at WOQ2620097 (FHLB’s Regulatory Plan for Globe (Sept. 22, 1989)). Those options were not available, however, in light of the then-depressed Oklahoma economy and financial stress on other Oklahoma financial institutions and sources of capital. Moreover, Globe’s options in raising capital were constrained by its circumstances. To preserve its NOLs, Globe would have had to raise capital without a “change of control.” Tr. 631:4 to 632:9 (Test, of Williams); Tr. 2915:19 to 2917:20 (Test, of O’Shaughnessy). An expert appearing for the government, Mr. David J. Ross, testified that capital could have been raised by issuing preferred stock or unsecured debt. Tr. 4425:16-24.

10 . During the trial, the government contested both the veracity of the testimony adduced by Globe regarding Ms. Fischman’s instructions about the interest-rate assumptions to be used in the projections and Ms. Fischman's authority to specify any such assumptions. See Tr. 3129:6 to 3135:20, 3139:14 to 3148:21, 3162:4 to 3192:14, 3213:20 to 3222:1 (Test, of Pittman). However, the court finds that Mr. Williams's testimony about the instructions was credible and that Ms. Fischman provided the instructions for the interest-rate assumptions to be used by Globe in performing the special analysis.

11 . In these factual circumstances, foreseeability does not constitute a limitation on awarding expectancy damages in the form of lost profits. Globe’s loss of profits was foreseeable on an objective basis to the government at the time it entered into the contract with Globe and Phoenix. As a comment to Restatement (Second) Contracts, puts it-

a. Requirement of foreseeability. A contracting party is generally expected to take account of those risks that are foreseeable at the time he makes the contract. He is not, however, liable in the event of breach for loss that he did not at the time of contracting have reason to foresee as a probable result of such breach. The mere circumstance that some loss was foreseeable, or even that some loss of the same general kind was foreseeable, will not suffice if the loss that actually occurred was not foreseeable. It is enough, however, that the loss was foreseeable as a probable, as distinguished from a necessary, result of his breach. Furthermore, the party in breach need not have made a "tacit agreement" to be liable for the loss. Nor must he have had the loss in mind when making the contract, for the test is an objective one based on what he had reason to foresee.

Restatement (Second) Contracts, § 351 cmt. a (emphasis added).

12 . More specifically, the government assumes that causation and forseeability could be established as part of a straightforward lost-profits case by Globe, but it seeks to shift the entire consideration of those elements to a focus on its hypothetical cost-of-replacement capital theory rather than on what Globe actually did to mitigate the breach:

Assuming plaintiffs believed that Globe would earn reasonably certain profits, the relevant question is what could the Government reasonably foresee plaintiffs would do in the event of a breach. The answer is: raise capital in order to pursue profitable opportunities. If Globe's risk-controlled arbitrage profits were as certain as Mr. Williams asserted ..., the only foreseeable result would be that plaintiffs would raise capital to pursue them____ Accordingly, plaintiffs’ failure to raise capital-the costs of which represent a cap upon damages-is the immediate cause of any lost profits.

Def.’s Reply at 14. This passage appears in a section of the government’s brief entitled, "Plaintiffs Failed To Demonstrate Causation Or Reasonable Foreseeability, And They Failed To Mitigate The Claimed Harm By Raising Capital.” Id. at 13.

13 . The government had initially retained Mr. Ross to prepare a report rebutting an early, now-abandoned claim by plaintiffs for damages based upon a hypothetical cost-of-replacement-capital measure of expectancy damages set out in an expert report prepared by Professor John J. McConnell. Tr. 4374:21 to 4375:7, 4332:13 to 4333:5 (Test, of Ross). This court granted the government summary judgment on that measure of damages because there was no dispute that Professor McConnell’s model was purely hypothetical and did not reflect actual events. See Globe Sav., 59 Fed.Cl at 96-97.

14 . There is evidence in the record that contemporaneously with the enactment of FIRREA Globe considered the possibility of raising capital either directly from its shareholders or by merging with a stronger thrift or bank. See supra, at 343 n. 9. In this respect, FHLB-Topeka's Regula-toiy Plan for Globe, dated September 22, 1989, reported that on July 11, 1989, Globe had submitted "a proposal for the voluntary liquidation of the institution dependent on whether certain issues are included [sic] in [FIRREA].” PX 296 at WOQ2620097. FHLB-Topeka indicated that several weeks later ‘‘Globe’s shareholders have informed our office that they [did] not intend to infuse additional capital.” Id. Finally, FHLB-Topeka related that "[m]anagement stated to our [o]ffice in a telephone conversation on August 25, 1989 that they are attempting to find an acquirer prior to submitting the application for voluntary liquidation.” Id. None of the capital-raising routes came to fruition for Globe, and it fell back on the plan to liquidate.

15 . This net interest income does not correlate directly to net income or lost profits because it does not take into account operating and overhead expenses on the one hand or revenue generated by reinvestment of cash received from runoff payments on the other.

16 . Mr. Davidson’s' liability-side proxy, duration-adjusted swap agreements, was keyed to "LI-BOR,” the London Inter-Bank Offered Rate. See PX 583 at 26 ("The use of the swap market proxy assumes that Globe's cost of funds was always equal to LIBOR.").

17 . Mr. Davidson calculated this spread on the basis of the spread derived from use of the Lehman mortgage index plus 50 basis points for a $700 million portfolio and the Lehman mortgage index plus 45 basis points for the $1 billion portfolio. Tr. 2422:8-24; PX 583 at 27-28.

18 . The government also argues that structural changes in the securities markets during the early 1990s diminished the viability of a risk-controlled arbitrage strategy and ultimately rendered it unworkable. Def.'s Br. at 13-15. In this respect, the government relies primarily on the testimony of one of its experts. Professor Frank Fabozzi. Professor Fabozzi did not make any analysis of Globe’s portfolio or of its detailed operating strategy. Tr. 3381:11 to 3384:12, 3390:20 to 3391:9, 3397:22 to 3398:9. He focused on institutions that had applied less discipline and fewer controls than had Globe in implementing such a strategy. Tr. 3489:18 to 3495:15. He also testified that the mortgage-backed securities market had become larger, more competitive, and more efficient during the 1990s: "there [may have been] opportunities, but they’re harder to come by, they’re not sustained that long, and they usually incur a risk that you probably didn’t expect." Tr. 3583:21-24. On behalf of Globe, Mr. Williams agreed with the statements, posed as leading questions, that "competition between investors will tend to produce an efficient market,” Tr. 1228:24 to 1229:12, and that "in a competitive market, ... prices rapidly impound any new information and make it difficult to consistently earn superior returns.” Tr. 1229:13 to 1230:8. Professor Fa-bozzi’s testimony was generally useful to the court, but it does not support the government's premise that Globe’s strategy was not viable in the 1990s, for four reasons. First, Globe had built a very substantial portfolio that carried substantially into the 1990s and in part throughout the 1990s, before running off to insignificance. Second, as both Mr. Williams and Mr. Davidson testified, a risk-controlled arbitrage strategy was being employed by many banks and thrifts at the time of trial in the fall of 2004. Tr. 71:18 to 72:19, 118:19 to 129:3 (Test, of Williams); Tr. 2260:10 to 2270:4 (Test, of Davidson). Third, Globe’s "business plan was not built upon acquiring under valued securities,” or trying to beat the market, but rather on trying to enter into appropriate transactions at the market. Tr. 1228:11 to 1232:21 (Test, of Williams). And, fourth, Mr. Davidson’s calculations based on the Lehman Mortgage Market Index show that even a non-selective strategy would have been viable and profitable, although not highly profitable, during the 1990s. See PX 583 at 30-31, Ex. K (Davidson Report).

19 . In addition, this greater emphasis on reverse repurchase agreements as a source of funds corresponds more closely to Globe’s actual liability composition as at June 30, 1989. See supra, at 337-38 & nn. 4-5.

20 . The seller identified as Midland Financial Corp. was the holding company for MidFirst, and the branch sold by Midland was the Stroud branch it had earlier acquired from Globe. Tr. 4021:4-8 (Test, of Bankhead).

21 . In Energy Capital, the court of appeals observed that "[ajlmost all of Energy Capital’s lost profits would have been earned after the date of judgment,” not before, so discounting was necessary in that case. 302 F.3d at 1330.

22 . Judgment will be entered only in favor of Globe. Phoenix was also a party to the contract with the government, but on the proofs of this case, it is entitled to no separate award of damages.