United States v. Davis, 803 F. Supp. 830 (S.D.N.Y. 1992)

U.S. District Court for the Southern District of New York - 803 F. Supp. 830 (S.D.N.Y. 1992)
October 27, 1992

803 F. Supp. 830 (1992)

UNITED STATES of America, Plaintiff,
George G. DAVIS, Gerald E. Lee, James H. Gilliland, P. Takis Veliotis and General Dynamics Corporation, Defendants.

No. 85 Civ. 6090 (KC).

United States District Court, S.D. New York.

Amended Opinion and Order October 6, 1992.

Second Amended Opinion and Order October 27, 1992.

*831 *832 Nancy Savitt, Beth Kaswan, David Koenigsberg, Asst. U.S. Attys., New York City, for plaintiff.

Nicholas Chabraja, Jenner & Block, Chicago, Ill., for defendants.


CONBOY, District Judge:

This is a lawsuit principally between the United States ("the Government") and one of the nation's leading defense contractors, General Dynamics Corporation ("GD"). The Government filed this case against GD in 1985, and the events that form the basis of this litigation occurred almost 20 years ago.

The Government's central claim in this case is that GD provided the Government with false ship-construction cost estimates when GD applied to the United States Maritime Administration ("MarAd") for ship construction subsidies known as Construction Differential Subsidies ("CDS"). GD sought these subsidies as part payment for GD's construction of two Liquid Natural Gas ("LNG") carriers for a partnership *833 known as the Lachmar partnership. The Government asserts that because MarAd relied on GD's false cost estimates when MarAd computed the size of GD's CDS's, GD is liable to the Government for damages, recision, and restitution. Accordingly, the Government is seeking in excess of $300 million from GD. The case was tried without a jury from December 10, 1991 to March 10, 1992.

GD maintains that the estimates it provided to MarAd in the 1970's were not fraudulent, but were fair and reasonable. The Government in substance claims that GD had internal, lower cost estimates which it did not disclose, and that, ipso facto, the estimates it did submit were not fair and reasonable. For the reasons that follow, we conclude that the Government has failed to carry its burden of proof, and enter judgment in favor of GD on all of the Government's claims. This shall constitute the Court's findings of fact and conclusions of law in this matter.

I. Statutory Background 
A. The Merchant Marine Act of 1936

The Merchant Marine Act of 1936, 46 App. U.S.C.A. § 1101 et seq., ("the act") as amended, establishes several programs to aid in the development and maintenance of a merchant marine in the United States. Two of the programs, one codified under Title V of the act and the other codified under Title XI of the act, are implicated in this case. We will discuss each program seriatim.

Title V of the act, better known as the Construction Differential Subsidy ("CDS") program, 46 App. U.S.C.A. §§ 1151-61, was intended by Congress to encourage ship purchasers, who wish to operate ships in foreign commerce, to place their orders for ship construction with American rather than foreign shipyards. Under the CDS program, when the Secretary of Transportation ("the secretary") is informed by a prospective ship purchaser that it wishes to have a particular ship constructed in an American shipyard, the secretary purchases the ship from the American shipyard at the ship's American price, and then sells the ship, at the vessel's lower foreign price, to the purchaser. By doing this, the secretary effectively subsidizes the purchaser's acquisition of the ship from the American shipyard. The Maritime Administration, which until 1981 functioned as part of the Department of Commerce and now is a component of the Department of Transportation, administers the CDS program.

Prior to 1970, the Secretary of Commerce[1] computed the proper size of a CDS by establishing, through competitive bidding, the domestic cost of building the ship in question, and then subtracting from that figure a fair and reasonable estimate of the cost of the ship if it were to be constructed in a foreign shipyard. 46 App. U.S.C.A. § 1152(a), (b). Under this system, a shipyard that created the design for the ship was at a disadvantage vis-a-vis other shipyards when competitively bidding for the ship's construction contract. This was so because when formulating its bid, the designing shipyard not only had to defray construction costs, but also had to defray the cost of creating the ship's design. Other shipyards were permitted to utilize the design specifications without cost when formulating contract bids, and were therefore able to underbid the designing shipyard. Thus, the pre-1970 statutory scheme created a disincentive to ship design. See Hearing on H.R. 15424 Before the Subcomm. on Merchant Marine of the House Comm. on Merchant Marine and Fisheries, 91st Cong., 2d Sess. 78-81 (April 9, 1970).

In 1970, Congress amended the act to do away with this disincentive. The amendment allows the Secretary to calculate CDS's using prices negotiated between ship purchasers and domestic shipyards as the domestic prices of the ships in question. Under this scheme, the designing shipyard is required to show its ship design only to a potential ship-purchaser, and not to potential *834 competitors. If the potential ship-purchaser wants to buy the ship, the designing shipyard and the purchaser establish the domestic construction price of the ship through negotiation. The ship-purchaser and the shipyard are then obligated to supply to MarAd back-up cost details and other evidence that shows that the negotiated price is "fair and reasonable." 46 App. U.S.C.A. § 1152(a) (ii).

Neither the act nor its legislative history elucidate the meaning of a "fair and reasonable" price in connection with a negotiated domestic price of a ship. However, when discussing the terms "fair and reasonable" in connection with estimated foreign prices of ships, Congress "recognized that the determination of estimated foreign costs is not an exact science and that reasonable men might differ over the estimate reached." S.Rep. No. 1080, 91st Cong., 2d Sess., reprinted in 1970 U.S.Code Cong. & Admin.News 4188, 4201. This observation applies as well to estimates of domestic costs of ships.

In 1971, MarAd issued proposed regulations which detailed the information that a shipyard must provide when submitting a CDS application to MarAd. Among other things, the proposed regulations required the shipyard to submit "[a] copy of all the detailed estimate backup sheets upon which the proposed price is based ... [and] historical actual cost labor data, as well as projected labor cost." JX132 at PO18494.[2] Moreover, the proposed regulations required that "[a]t the conclusion [] of negotiations, the shipyard shall certify that the cost or pricing data which it has furnished is current, complete, and accurate." JX132 at PO18495 (emphasis added). These regulations were, however, never enacted.

In 1972, MarAd proposed a second set of regulations detailing the information a shipyard must provide to MarAd when submitting an application for a CDS. Like the 1971 regulations, the 1972 regulations required the shipyard to submit historical as well as projected labor costs. Construction Differential Subsidy, 37 Fed.Reg. 6759, 6763 (1972). Instead of the certification requirement contained in the 1971 regulations, the 1972 regulations mandated that the "applicant shall file ... any amendments necessary to keep all information contained or furnished in connection with a pending application current and correct." Id. (emphasis added). Like the 1971 regulations, the 1972 regulations were never enacted. Regrettably, the Government has been unable to provide any internal documentation from the agency that would illuminate the purpose, significance, or relevance to the bureaucratic process of these proposed changes.

MarAd's Division of Domestic Costs ("the Division") had the responsibility of reviewing the shipyard's backup cost details, conducting negotiations with the shipyard, and then determining whether or not to recommend to the Maritime Subsidy Board that the resulting negotiated domestic ship price was fair and reasonable. The Division of Domestic Costs had no authority to accept a proposed domestic price on behalf of MarAd. Only the Subsidy Board had the authority to enter into contracts for the award of CDS's.

Under Title XI of the Act, a ship-purchaser can finance its purchase of an American made ship by selling bonds and using the bonds' proceeds to pay for the ship's construction. "The bond[s] [are] guaranteed by the Government, and in the event of default, the bondholder[s] may pursue the Government for [the bondholders'] claim." Liberty Maritime Corp. v. United States, 928 F.2d 413, 415 (D.C.Cir.1991). Using this scheme, the Government "encourage[s] private investment in the production and rebuilding of ships [in American shipyards]." Id.

B. The Truth in Negotiations Act

The Truth in Negotiations Act ("TINA"), whose predecessor statute Congress enacted in 1948 to ensure, inter alia, that the prices the Government pays for goods it buys for the Department of Defense are fair, reasonable, and economical, see 10 *835 U.S.C.A. § 2301(a) (1), is not directly implicated in this case. However, given the absence of case law that discusses what type of cost information the Merchant Marine Act requires a shipyard to provide to MarAd, the parties and we find it useful to examine TINA's reporting requirements.

TINA requires a Government contractor to disclose to the Government only those

facts that, as of the date of the agreement on price of a contract (or the price of a contract modification), a prudent buyer or seller would reasonably expect to affect price negotiations significantly. Such term does not include information that is judgmental, but does include the factual information from which a judgment was derived.

10 U.S.C.A. § 2306a(g) (emphasis added). Thus, TINA only requires a contractor to provide factual information that can be verified to the Government, not judgments based on factual information. See id.; Inter-Con Sec. Systems, Inc., 84-2 BCA ¶ 17,274 (1984) (TINA case recognizing distinction between fact and judgment).

II. Historical Background 
A. QSD Management

In 1964, GD purchased the Quincy Shipyard from Bethlehem Steel, and it became GD's Quincy Shipyard Division ("QSD"). Initially, QSD worked on Navy contracts for ships and nuclear submarines, but by 1971 QSD had completed most of the work it had on order and was searching for new business. P. Takis Veliotis served as QSD's General Manager from January 1973, until October 1977, and reported directly to David Lewis, then General Dynamics' Chairman and Chief Executive Officer.

GD required its divisions to prepare operating plans annually. In the fourth quarter of each year, QSD submitted a preliminary operating plan for the following year, which, after a process of review and revision involving both General Dynamics and QSD management, was approved early in the following plan year.

QSD's operating plans projected expenses and revenues for the plan year as well as for three years forward, and each quarter QSD evaluated its performance by comparison to the plan. GD433-441[3]; JX311. It is plain and undisputed that QSD routinely failed to achieve the targets set in the operating plans. In particular, and critical to the events underlying this lawsuit, QSD significantly underestimated costs at completion in its 1974, 1975, 1976 and 1977 operating plans. GD467; Tr. 3531-38 (Murphy, 2/18); Tr. 2845-47 (Beggs, 2/6).[4] Arthur Andersen & Co. ("Arthur Andersen") served as GD's independent outside auditor and performed quarterly reviews of GD's divisions, including QSD. It also reviewed QSD and all other GD divisions as part of its annual audit of GD's consolidated financial statements.

During the 1970's, LNG ship construction became the bulk of QSD's work. JX716 at 9 (Lewis dep.); JX717 at 3-4 (Lewis Dep.); Tr. 506-08 (Whittemore, 12/ 13). After extensive study, QSD chose the spherical Moss-Rosenberg freestanding tank design for its LNG ships and became the only United States licensee of that design.

On September 28, 1972, QSD signed contracts to construct three LNG ships, the first it built, for companies related to Burmah Oil ("Burmah"). JX07; JX08; JX09; JX698 at 12198; JX515 at 0554, 0557; Tr. 1023-24 (McGowan, 12/19). QSD's contracts for the Burmah ships called for the first ship to be delivered in December 1975, the second in March 1976, and the third in March 1977. JX12 at 10412; JX698 at 12180; JX515 at 0557. The contract price for each of the Burmah ships was $89.02 million, subject to certain minor adjustments, and the contracts were "firm fixed-price" contracts that did not provide for any adjustment due to escalation. JX07; JX08; JX09.

*836 Each of the purchasers of the Burmah ships applied to MarAd for a CDS, and QSD supplied MarAd with "proposal support" to demonstrate the fairness and reasonableness of the proposed price. JX04; Tr. 1023-24 (McGowan, 12/19). The 1972 proposal support included an estimate of the amount it would cost to build the ships. Most significantly, with respect to labor hours, QSD estimated that it would expend 2.482 million recurring operations labor hours building the first ship, 2.283 million hours building the second, and 2.175 million hours building the third. JX04 at MAR04 0165.

Because QSD lacked experience in both building LNG tanks and welding thick-plate aluminum, QSD hired Pittsburgh-Des Moines Steel Company ("PDM") to build the spheres using the Moss-Rosenberg design. GD72; JX698 at 12179; JX515 at 0556; JX716 at 20-21 (Lewis Dep.); Tr. 2838 (Beggs, 2/6). PDM had a construction facility in Charleston, South Carolina.

In 1973 and 1974, QSD entered into contracts to build five additional LNG ships, often called the "Cherokee" ships. MarAd did not subsidize these ships through the CDS program, but it did grant loan guarantees under Title XI of the Merchant Marine Act. Ultimately, the first of these five ships became the third LNG ship built by QSD and the third Burmah ship became the fourth ship in the series. JX698 at 12178; Tr. 2556-57, 2610 (Fisher, 2/4); Tr. 4016 (Murphy, 2/20).

PDM encountered substantial difficulty constructing the first LNG ships, and in December 1974, QSD was forced to terminate PDM's contract, stop work, and take over the Charleston facility. JX698 at 12180; GD96; JX530 at 9469; JX526; Tr. 1020-22 (McGowan, 12/19); Tr. 1079 (McGowan, 12/20).

Although QSD invested over $50 million to improve the Charleston facility, it too had substantial difficulties building the spheres. In November 1975, QSD General Manager Veliotis told the GD Board of Directors that the first sphere would be completed by February 1976. By the summer of 1976, however, QSD was not even close to completing the first sphere. On June 30, 1976, MarAd's on-site construction representative reported that the situation at Charleston was critical and growing more serious each week. JX61 at 136093; JX549 at 14701; Tr. 1093-94 (McGowan, 12/20). QSD's problems were not limited to sphere production at Charleston. QSD encountered large increases in the number of labor hours needed to build hulls at Quincy. Moreover, QSD also experienced dramatic increases in its material costs as a result of inflation, volatility of economic indicators, severe shortages in supplies of materials, failure of certain vendors to honor firm price contracts, and greatly increased steel scrap rates. JX04 at MAR04-0128; JX196 at 102911; Tr. 3481-82 (Murphy, 2/14); Tr. 3518-22 (Murphy, 2/18; Tr. 2758-59 (Parsons, 2/5).

These problems resulted in large cost overruns on the first ships. With virtually every estimate QSD prepared, its projected costs at completion increased. By May 1975, the estimated cost at completion ("CAC") of the first ship was $125.2 million. The 1976 operating plan projected a CAC for that ship of $140.6 million, and an August 1976 status report projected a CAC of $155.1 million. JX645 at 121548, 121550; JX750 at 120584; JX63.

In 1976, QSD executed agreements to build two additional LNG ships, which were expected to be the ninth and tenth in the series. These ships were often called the "Cherokee VI" and the "Cherokee VII" ships. MarAd did not subsidize these ships and the selling price was negotiated solely in the marketplace. Significantly, QSD and the purchaser agreed on a fixed, delivered price of approximately $155 million per ship.

B. The Lachmar Partnership

On May 7, 1976, GD, Panhandle Eastern Pipeline Corporation, and Moore McCormack Lines, Inc., acting through wholly-owned subsidiaries, formed the Lachmar partnership ("Lachmar"). Pursuant to certain agreements, QSD assumed the responsibility for constructing the ships, MooreMcCormack *837 for operating them, and Panhandle for shipping the LNG. JX758.

As set forth in greater detail below, the construction contracts required QSD to construct two LNG ships substantially in accordance with GD's then-current standard specification, subject to changes as permitted by the contracts. The base price of March 30, 1976 was $122,538,000 per ship. Ninety percent of the price was subject to escalation according to various formulas set forth in the contract to reflect future increases in the costs of material and labor. The base contract price also was subject to adjustment to reflect any added cost incurred as the result of changes requested by the partnership or required by regulatory authorities. The construction contracts specified delivery dates for the ships of December 4, 1979, and March 18, 1980. These dates were compatible with the anticipated date of the commencement of deliveries of LNG. JX150 at XXXXXX-XX; JX758 at 2-4; JX565 at MAR09 0434.

The Lachmar Partnership Agreement contemplated that Lachmar would finance 25% of Lachmar's actual cost for the two LNG ships with capital contributions from the partners, subject to a cap of $115 million. Lachmar intended to finance the balance of its cost for the LNG ships by issuing and selling Title XI bonds as provided by the Merchant Marine Act.

The subsidy applications for these two ships are at the heart of this case.

C. The Negotiations with MarAd

On June 8, 1976, a meeting was held between representatives of Lachmar and high level MarAd officials. JX156; JX158. The purpose of the meeting was to provide MarAd with an overview of the Lachmar project and to discuss scheduling with MarAd. Veliotis indicated that "detailed cost information would be provided by the shipyard at about the time of the filing of the application. Preparation of this information is currently underway." JX158 at MAR05 1963; JX156 at XXXXXX-XX.

On June 22, 1976, Lachmar submitted to MarAd its applications for CDS's and for Title XI guarantees. On July 19, 1976, representatives of QSD and MarAd met in Washington, D.C. to discuss the proposed domestic purchase price of the Lachmar ships, and more specifically, the escalation clauses of QSD's construction contracts with Lachmar. At this meeting, QSD submitted to MarAd its "Proposal Support Characteristics Information" ("Proposal Support"). This document supported a cost breakdown of the $122.5 million base contract price per ship. JX180.

On July 22, 1976, MarAd sent QSD a letter requesting further estimating details, quotations, and historical cost information. On August 11, 1976, QSD responded by providing MarAd with a packet of detailed cost estimating information and historical cost data. JX182; JX196. During the negotiations that followed, MarAd was represented primarily by John McGowan who had been working for MarAd since 1939, and had been Chief of the Division of Domestic Costs since 1967. Up to the time of the negotiation, McGowan had devoted his entire professional life to ship cost estimating and had been involved in all cost estimating undertaken by MarAd between 1967 and 1977 relating to new ship construction. McGowan's department also controlled progress payments, which involved assessments by MarAd and its representatives of the extent of progress a shipyard had achieved in building a ship. Furthermore, McGowan was involved with change order and mortgage issues, both of which encompassed cost estimating. Tr. 909-10, 976-81 (McGowan, 12/19).

McGowan's principal assistant was David Gessow. Among other things, Gessow took comprehensive notes during the negotiations and compiled a detailed record of MarAd's review of QSD's proposal support, the August 11 estimating detail, and other information. QSD was represented at the negotiations by Gary Grimes, George McAndrew, and Robert Renn. JX611/GD403 at 0847; JX618/GD404 at 0844; JX708 at 117-28 (Gessow dep.); Tr. 923 (McGowan, 12/19).

In July 1976, MarAd asked the Defense Contract Audit Agency ("DCAA") to conduct *838 a limited audit consisting of a review of labor and overhead rates at QSD's Quincy shipyard and Charleston facility to assess whether those rates, both currently and as projected, were reasonable. JX181; Tr. 935-36 (McGowan, 12/19). On September 27, 1976, the DCAA submitted its audit report to MarAd and informed MarAd that the cost or pricing data as submitted by QSD was acceptable as a basis for negotiating a price. JX215.

A few days earlier, at a meeting on September 21, MarAd had informed QSD that it considered QSD's escalation clause unacceptable. MarAd demanded: (1) that QSD remove fixed depreciation from the escalatable base; (2) that QSD substitute index-based escalation for "pass-through" escalation that QSD had proposed for Attachment A items; (3) that QSD eliminate the Quincy and Charleston wage rates as a basis for computing labor escalation, and instead use a national Bureau of Labor Statistics ("BLS") index; (4) that QSD allocate escalation by element by quarter according to a proposed construction schedule rather than actual construction progress. QSD responded that MarAd was asking it to assume risk, which might require an increase in the base price of the ships, and MarAd acknowledged that a price increase might be necessary.

QSD submitted to MarAd a "Supplemental Pricing Summary" dated October 1, 1976, which revised QSD's proposal support so that it would comply more fully with MarAd's guidelines, including those regarding escalation. The total base price proposed by QSD in its October 1, 1976 submission was $127.5 million per ship. This proposed base price included a 10% increase in the estimated base cost of attachment A items and an 8.5% increase in the estimated cost of Quincy labor, in order to compensate for the increased risk inherent in MarAd's proposed escalation provisions. JX221. Ultimately, Lachmar and MarAd both agreed to an increase in the base price of each ship to compensate QSD for the additional risk it would have to assume under MarAd's escalation clause. JX253, Tr. 931-33 (McGowan, 12/19).

Between October 14 and November 18, 1976, QSD and MarAd met repeatedly to discuss the contents of QSD's proposal support and estimating detail. JX609/GD402; JX708 at 117-18 (Gessow dep.). During a meeting on November 17, 1976, QSD advised MarAd that QSD recently had reached agreement with Lachmar as to the amount by which the base price of the ships would be increased as a result of the substitution of MarAd's escalation provision for that of QSD. QSD told MarAd that it and Lachmar had agreed to a revised base price of $126.5 million. JX609/GD402 at 0919. After conferring among themselves, MarAd's representatives told QSD that its price of $126.5 million was about $1 million too high, and that by lowering its price by $1 million, QSD could resolve all the open issues under discussion. JX609/GD402 at 0919; Tr. 1145-48 (McGowan, 12/20).

The following day, November 18, 1976, QSD offered to reduce its base price to 125.9 million to compromise all open issues. MarAd's representatives accepted that price. The parties turned their attention to the estimated delivered price and agreed on an estimated delivered price of $155 million per ship. JX609/GD402 at 0920-22; Tr. 1151-55 (McGowan, 12/20).

On November 23, 1976, Lachmar and MarAd agreed on an estimated foreign price of $115.5 million per ship. Thus, the resulting CDS rate for the Lachmar ships was 25.48%. GD219; JX374 at 5. A number of matters unrelated to the price of the LNG ships, such as FPC (Federal Power Commission) approval of a related transportation agreement entered into by Lachmar, delayed the Maritime Subsidy Board's approval of the CDS until July 26, 1977. On that day, the Maritime Subsidy Board approved for the Lachmar ships: (1) the negotiated base price of $125,919,000; (2) the estimated delivered price of $155,000,000; (3) a foreign price of $115,500,000; and (4) the CDS rate of 25.48%. JX371; JX374. MarAd notified Lachmar and QSD of the Subsidy Board's approval on that date.

*839 III. GD's Establishment of the Ship Price

In late 1974, QSD developed a market price for an LNG ship of $110 million, subject to escalation from October 1974 to the quarter in which the ship would be delivered. JX747 at D008382. The $110 million price initially was linked to a summary level cost estimate as of October 1, 1974, and there is no evidence in the record of any cost estimating detail to support the cost breakdown in the October 1974 estimate.

When QSD negotiated the price for two LNG ships with the Lachmar partnership in the spring of 1976, it chose simply to use the October 1974 market selling price of $110 million, with escalation forward from October 1974 to the date of delivery. JX747 at D008382; JX597 at 104539.

Before escalating the $110 million price to March 31, 1976 in order to obtain a base price, QSD made two adjustments. First with Lachmar's approval, it raised the price slightly from $110 million to $111.4 million. Second, it reallocated the entire $111.4 million price into four cost categories: certain identified components known as "Attachment C" items (later called "Attachment A" items) (10%); other material (41%); Quincy labor (42%); and Charleston labor (7%). These categories did not correspond to the cost elements in the October 1974 estimate. At that time, "Charleston labor" was entirely a subcontract (and thus a material) item. JX598 at XXXXXX-XX; JX107; JX747 at D008384.

In escalating the October 1974 price to the current period, QSD sought to choose an escalation model that would adequately measure the increases in prices that had occurred during that time. The escalation exercise simply sought to preserve whatever margin actually was present in the October, 1974 price, and to ensure that the escalation in prices did not erode that margin. JX75; Tr. 169-71 (Renn, 12/11); Tr. 370-71 (Renn, 12/12).

QSD applied different escalation factors to each of the four cost categories to bring the October 1974 selling price forward to the current period. The result was a new April 1, 1976 base price of $121,538,000. JX598 at 104630. This price, which was competitive in the marketplace, pleased the other Lachmar partners, and was less than what the Lachmar partners had estimated it would be. JX700 at 44 (Campbell dep.); Tr. 2548-71 (Fisher, 2/4).

During the negotiations between Lachmar and QSD, Lachmar agreed to a further increase in the base price of one million dollars to protect the delivery slots that Lachmar needed. The base price was thus increased to $122.5. JX724 at 36 (Olfson dep.); GD132; GD134 at 133233; JX124 at 101648. The parties also discussed the manner in which the base price would escalate in the future. QSD sought to escalate the entire base price, but the other Lachmar partners wanted 15% of the price to remain unescalated. The parties ultimately compromised that 10% of the base price would remain fixed. JX99; JX124 at 101656

Having reached agreement on the fixed portion of the contract price, the parties allocated the remaining 90% of the base price back among the four major cost elements of the project, in roughly the same proportion that those cost elements previously had borne to each other. The parties also agreed on the method in which each of these cost elements would escalate in the future. JX124 at XXXXXX-XX.

IV. Development of Proposal Support

In May 1976, after executing the construction contracts with Lachmar, QSD began to assemble evidence to submit to MarAd that the $122.5 million base price was fair and reasonable. JX132; JX760. In seeking to demonstrate to MarAd that the basic cost allocations in the construction contract were reasonable, QSD did not use or provide to the Government the cost estimate it had prepared almost two years earlier, in October, 1974. Rather, QSD assembled evidence of the reasonableness of the price based on the shipyard's actual construction experience through the first half of 1976. JX180; JX196; JX172; GD27.

*840 This exercise was consistent with advice that QSD received from its lawyers. On May 17, 1976, Clay Cook, an attorney for QSD from the firm of Cadwalader, Wickersham & Taft ("Cadwalader") met with Donald Frye of MarAd. Frye provided Cook with a copy of regulations that MarAd had proposed in 1971, but never enacted, and with a copy of regulations MarAd had proposed in 1972, but also never enacted. JX132. In addition, Rona Goodman, QSD's in-house counsel, contacted James Rossi of Cadwalader to solicit his suggestions regarding what QSD should submit to MarAd. Rossi essentially recommended that QSD review its previous submissions to MarAd and reconcile those submissions with the price QSD proposed for the Lachmar ships. JX760.

Several days later, QSD developed the "LNG Master Planning Assignment," which had several functions. One part of the assignment was to forecast a cost at completion for ships 1-8 and for ships 9-12 as part of the development of the QSD's 1977 operating plan. JX133; JX171 at 109720. Second, the document established tasks to be performed in connection with the preparation of the proposal support. One task was to consider different ways to absorb the difference between the assignment's operating plan forecasts at completion, on the one hand, and the cost allocations set forth in the Lachmar construction contracts on the other. Another task was to prepare the reconciliation suggested by Rossi. JX133; JX171 at 109720.

QSD's Financial Analysis Department ("Financial Analysis") began assembling cost estimates that supported the negotiated Lachmar base price of $122.5 million. Because QSD was preparing cost estimating detail for MarAd to support the reasonableness of previously selected cost allocations, QSD conducted various mathematical exercises to assess different combinations of estimating parameters that would correspond to the summary level cost allocations in the construction contracts. Tr. 196-97 (Renn, 12/11).

V. Truthfulness of the Construction Contract

We address at the outset the Government's contention that the following statement in the construction contracts, copies of which were appended to the Lachmar CDS application, is false: "The Base Contract Price has been determined on the basis of estimates as of March 30, 1976, for costs of work to be performed by subcontractors, materials, machinery, equipment, supplies, labor, and overhead." The Government suggests that McGowan was misled by this sentence into believing that QSD had prepared its detailed cost estimate before the price was established with Lachmar. JX158; Tr. 991-93 (McGowan, 12/19).

It is apparent that the challenged statement did not have the effect suggested by the Government. McGowan and others at MarAd knew that QSD prepared its detailed cost estimate after QSD negotiated the price with Lachmar. JX158; Tr. 991-93. In addition, on June 8, 1976, Veliotis and representatives of Lachmar met with Robert Blackwell, Tom Pross, who was McGowan's immediate supervisor, and other senior MarAd officials. Veliotis indicated that detailed cost information would be provided by the shipyard at about the time Lachmar filed its CDS application, and that the preparation of the cost information was currently underway. These facts are reflected in a memorandum concerning the meeting that Pross prepared for the file and sent to McGowan. See JX158. (McGowan, 12/19).

Thus, McGowan knew from the memorandum that, notwithstanding that the parties had agreed to the base price of the ships, the shipyard had not yet prepared its detailed cost support. McGowan was not aware of the means by which QSD and Lachmar had arrived at the price of the ship. What he knew was that the price had been determined and the yard was going to develop its proposal support, i.e., its detailed estimate. Tr. 991-93 (McGowan, 12/19). Thus, there is no evidence in the record that McGowan or anybody else at MarAd was misled by, or relied in any way *841 upon the statement in the construction contracts challenged by the Government.

In any event, the statement in the construction contracts is not false. As set forth above, the $122,538,000 base price was determined on the basis of cost estimates for each of the four categories of work, escalated to March 30, 1976. The construction contracts do not represent that a detailed, "bottoms-up" cost estimate preceded the development of the base price. Tr. 3619-24 (Murphy, 2/18).

VI. GD's Pricing Detail Estimates

We observe at the outset that there were in the spring and summer of 1976 a variety of circumstances that affected the viability of estimates being made in connection with the LNG program. For any estimate of costs for a future production event that unfolds sequentially and progressively, such as the construction of an LNG ship, there is a range of outcomes that is reasonably likely. Tr. 3517-18 (Murphy, 2/18); Tr. 1755-56 (Rylander, 1/9). The breadth of the range is dictated by the amount of risk or uncertainty that is present. Tr. 3517-18 (Murphy, 2/18); Tr. 1752, 1820-21 (Rylander, 1/9).

In the summer of 1976, when QSD attempted to demonstrate the reasonableness of the Lachmar base price, there was considerable risk and uncertainty concerning QSD's first LNG ship, to say nothing of the amount of risk and uncertainty concerning the eleventh and twelfth ships. As of 1976, GD stood to lose millions of dollars on the LNG ships then under construction, and had not yet established any record of successful or profitable construction. JX697.

As detailed further below, not even the first sphere, much less the first ship, had been completed. Moreover, the prognosis for sphere production at that time was bleak. See JX548. QSD's effort to construct the hulls at Quincy was no more successful. The shipyard was experiencing substantial overruns of its initial budget estimates in the number of manhours needed to build the first hulls under construction. Moreover, during this same time frame, QSD was witnessing continuing deterioration in the labor hour estimates at completion for the first ships; as virtually every new estimate was prepared, the number of hours projected at completion increased. Nothing was fixed. GD366 at 14081; JX04 at MAR04 106; JX180 at 15; Tr. 1028-39 (McGowan, 12/19); Tr. 2845-46 (Beggs, 2/6)

QSD's difficulties were exacerbated by a history of labor problems and unrest at the shipyard. Tr. 3526-27 (Murphy, 2/18); Tr. 2835-36 (Beggs, 2/6). Moreover, the construction of LNG ships presented unique and complex problems for the shipbuilding industry. In the summer of 1976, no American shipyard had ever successfully built an LNG ship. Indeed, no LNG ship in the world the size of that built by QSD had ever yet successfully carried a cargo of LNG. The construction effort undertaken by QSD was new, still untested, and fraught with uncertainty and risk. Tr. 1019 (McGowan, 12/19); Tr. 2537, 2561, 2575-78, 2581 (Fisher, 2/4); Tr. 2479-80, 2484-88 (Glasfeld, 2/4).

Finally, the volatility and instability of the national economic environment in the 1970's made cost estimating difficult. Tr. 3517-18 (Murphy, 2/18).

We will now review the details of GD's component elements in its overall estimate: A. the material cost estimate; B. the sphere labor estimate; C. the operations recurring hours estimate; D. the operations non-recurring hours estimate; E. the engineering hours estimate; and F. the sphere insulation and joinder estimate.

A. The Material Cost Estimate

There are two basic elements to the material cost estimate that QSD submitted to MarAd to support the reasonableness of the Lachmar price: (1) a base cost estimate as of March 1976; and (2) escalation. JX180. QSD developed bills of material (or estimates of quantity) based on drawings, design specifications, and usage estimates; it obtained quotations from vendors as of March 30, 1976, for all components with a per/ship value of $1,000 or more; it based steel costs on actual March 1976 steel *842 prices; and it based aluminum costs on a current Alcoa quote as of March 30, 1976. JX196 at 102900.

QSD proposed two types of escalation of the material base cost. First, with regard to certain components identified as "Attachment A" items, QSD proposed that the actual cost would be passed on to the purchaser, and that QSD would earn no profit since it assumed no risk. Second, QSD proposed that the remainder of the material estimate escalate in accordance with the movement of selected national indices. JX180 at 33-35.

MarAd flatly rejected the shipyard's proposal to pass through the actual costs of the attachment A items, and insisted that all material escalate in accordance with national indices. QSD then requested and received a ten percent increase in the base cost of the attachment A items, to compensate QSD for the possibility, supported by historical evidence, that the national indices would not track adequately increases in the cost of the specialized Attachment A components. This increase was referred to as a "factor for index shortfall." JX222; GD202; Tr. 3495-502 (Murphy, 2/14).

Rylander, the Government's expert at trial, conceded that he found no misrepresentation with respect to any element of the material base cost estimate, with the exception of the steel scrap rate. Tr. 2079, 2091 (Rylander, 1/28); Tr. 2115 (Rylander, 1/29). Murphy, GD's expert at trial, testified, and the Court finds, that there was no fraud or misrepresentation in connection with the bills of material, quotations, or any other aspect of the material base cost estimate. Tr. 3471-75 (Murphy, 2/14). Moreover, there was no fraud or misrepresentation in the 25% steel usage factor used in the proposal support. As QSD demonstrated to MarAd, steel scrap rates on LNG ships under construction ranged as high as 30% and the aggregate usage factor for all ships under construction through June 1976 was 25%. Murphy confirmed the accuracy of this data, JX196 at XXXXXX-XX; Tr. 3484 (Murphy, 2/14), and we find unpersuasive Rylander's conclusion that QSD should have used a scrap rate of 15%. In sum, we find that the base material estimate that GD provided to MarAd was fair and reasonable.

Apart from the base cost submittal, the only remaining aspect of the material estimate pertains to escalation. Rylander again conceded that he found no fraud or misrepresentation in connection with the shipyard's escalation proposal. JX180 at 33-35; Tr. 2101 (Rylander, 1/28). Moreover, there also was no fraud or misrepresentation in QSD's request for a factor for index shortfall on the attachment A items. Murphy reviewed the factual data that QSD submitted in connection with its request and confirmed that the data accurately showed that, in the past, prices for the Attachment A items had risen faster than the general material indices. Rylander conceded that he did not find any false statement in connection with QSD's submittal. GD202; Tr. 3495-502 (Murphy, 2/14); Tr. 2401-02 (Rylander, 1/30).

At trial, the Government suggested that QSD overstated the historical increase in the Attachment A items because some of the earlier prices for those components had been negotiated in 1973 for delivery in 1974. Murphy explained that the prices still can be viewed as 1974 prices, and we so view them. In any event, the evidence showed a 28% index shortfall in the Attachment A items, but QSD only requested a 10% increase in the base price. Tr. 4149-57 (Murphy, 2/21).

Although he found no misrepresentation in connection with either the base cost estimate or the escalation proposal, Rylander asserted that the Government had been defrauded by almost $10 million because he found an internal QSD estimate (prepared by Walter Sullivan) of escalated or "at-delivery" material costs that was different from the escalated material estimate agreed to by QSD and MarAd. This internal estimate was incorporated into QSD's operating plan, which, as we have indicated, was almost always shown by later developments to be way off the mark, reflecting more division boosterism than cleareyed reality. JX178 at 113410; Tr. 2101-02, 2074-76 (Rylander, 1/28).

*843 Rylander engaged in a false comparison. Most importantly, he was unable to establish that the two estimates used the same escalation assumptions. Rylander conceded that he did not know how Sullivan had escalated the material costs in his analysis, and Rylander also conceded that MarAd's escalation formula would be different. Tr. 2085-86; 2090 (Rylander, 1/28).

Rylander's comparison is also flawed because the Sullivan analysis may not have included certain material cost contingencies that were included in the proposal support and openly disclosed to MarAd. Rylander again conceded that he did not know the extent to which these identified contingencies had been excluded from the Sullivan analysis. Tr. 2092-94 (Rylander, 1/28).

At bottom, Rylander acknowledged that all he did was find a couple of documents that differed, and concluded that there was fraud because the two documents did not contain the same numbers. Moreover, he admitted that he found nothing to establish that, as of November 1976, the shipyards internal forecast was in any way guaranteed. The Government ultimately contends that QSD committed fraud simply by negotiating a contract that provided protection against potential future inflation in the national economy that QSD apparently hoped it could avoid. Indeed, Rylander went so far as to testify that QSD should have requested a downward adjustment in the material estimate because it hoped to experience less inflation than the parties estimated would be measured by the national indices demanded by MarAd. Tr. 2096, 2102 (Rylander, 1/28); Tr. 2401, 2405-06, 2411 (Rylander, 1/30).

In light of the rampant inflation that had recently occurred and that QSD had experienced on the LNG ships under construction, QSD did not commit a fraud by requesting an escalatable contract without a downward adjustment in the indices chosen by MarAd. The type of contract dictated by MarAd, in other words, a contract that was not a "cost-plus" contract, specifically permitted a shipyard to earn additional profit if it held costs below the amount MarAd found to be fair and reasonable, just as the shipyard bore the risk of all cost overruns. There was no fraud in QSD's material estimate. GD468; Tr. 1005-08 (McGowan, 12/19); Tr. 3503 (Murphy, 2/14); Tr. 3521-24 (Murphy, 2/18).

If MarAd had accepted QSD's original escalation proposal, it would have ended up paying $13.2 million less than the amount it ultimately paid. QSD's original escalation proposal, in which it was willing to pass along its actual incurred cost and forego all profit on the Attachment A items, is inconsistent with the actions of a party intent on defrauding the Government. GD466; Tr. 3503-10 (Murphy, 2/14).

The Government seeks additional damages in the material area based on two separate theories that relate to events that occurred after the agreement on price. The first theory is based on another comparison that Rylander made, this time between the escalated material estimate agreed to by QSD and MarAd on the one hand, and a new escalated price estimate that Sullivan prepared in 1977, on the other. This theory suffers from the same infirmities as Rylander's comparison to the 1976 Sullivan estimate, with the additional problem that the 1977 estimate was never even accepted by QSD management for operating plan purposes. JX327; Tr. 2145-49 (Rylander, 1/29).

The Government's second theory is based on the fact that, well after the parties reached agreement on price in November 1976, but before the parties signed the construction contracts in July 1977, QSD began securing more favorable material purchase commitments. In many instances, QSD was able to secure commitments to purchase material at prices less than the escalated material estimate agreed to by QSD and MarAd. QSD did not volunteer this fact to MarAd during the months in which execution of the contract was delayed. The delay was for reasons unrelated to the agreement on price. Tr. 937 (McGowan, 12/19); Tr. 3476-81 (Murphy, 2/14).

The record indicates that QSD had no options or precommitments to purchase material prior to the agreement on price in *844 November 1976. Tr. 3490 (Murphy, 2/14); Tr. 2158 (Rylander, 1/29). MarAd obviously was aware that after the agreement on price, QSD was continuing to build LNG ships and was going to generate a tremendous volume of additional data, but there is no evidence that MarAd ever requested any updated information prior to the execution of the contracts.

Indeed, McGowan told QSD that if it wanted to go ahead and purchase material in advance of the Contract award, QSD bore all the risk of doing so. Tr. 1015 (McGowan, 12/19). Moreover, we find that it is a common and prudent business practice to purchase material after agreement on price and before contract execution. Tr. 3476-81 (Murphy, 2/14).

The Government maintains that QSD created an obligation to disclose the 1977 material purchase commitments because of a July 18, 1977 meeting at which McAndrew told McGowan that some of the cost items QSD had proposed had already been or were about to be exceeded. Specifically, McAndrew apparently gave McGowan information relating to the increase in hours on the first ship from 3.8 million hours to 3.95 million hours. These statements were truthful, and there is no evidence that they had anything to do with QSD's material costs. JX366; Tr. 2320 (Rylander, 1/30).

The Government ultimately argues that McAndrew's statements caused MarAd to decide not to reopen the Lachmar price negotiations despite the potential cancellation of the Cherokee VI and VII ships. However, McAndrew apparently made these statements immediately after a discussion with McGowan about the weighting of work for progress payment purposes, and before a discussion about labor problems at QSD and Veliotis' attempts to get value for hours worked. McAndrew's statements about increased labor costs are directly relevant to that latter discussion. JX366. In any event, it does not appear from the memorandum that the statements were made to persuade MarAd not to reopen negotiations in light of the potential cancellation of the Cherokee VI and VII ships. That issue arose later in the meeting; one of the concluding paragraphs of the memorandum notes that "[McGowan] then spoke of some other modifications [to the contracts] ... which had to do with some items [Grimes] and [McGowan] had talked about viz resequencing ships and effect on material and labor costs...." It does not appear from the memorandum that McAndrew even addressed the issue, other than to urge McGowan to "please get [his other modifications] on the table because [the parties] were coming down to the wire." JX366.

In sum, the statements of McAndrew challenged by the Government were not false, and there is no evidence that they were incomplete or misleading in any way in the context in which they were made. JX366.

Moreover, the Government has also failed to show its detrimental reliance on McAndrew's statements. McGowan, who heard the statements, nevertheless continued to believe that a clause should be added to the Lachmar subsidy contracts to provide for an adjustment to the escalation tables in the event the Cherokee ships were cancelled. He strongly urged that position on his superiors, including John Nachtsheim. JX379; Tr. 954-59 (McGowan, 12/19).

For its part, QSD was willing to adjust the escalation tables. However, it contended that, if the price of the Lachmar ships was reopened to account for potential cost savings resulting from the cancellation of the Cherokee ships, the parties should consider potential cost increases also resulting from the cancellation (such as those resulting from the fact that fewer ships would absorb the overhead pool, or that the Lachmar ships would experience higher labor hours as a result of moving back up the learning curve). For its part, Lachmar opposed any reopening of the price, because it would not accept early delivery of its LNG ships even if the Cherokee ships were cancelled. JX379; Tr. 943, 954-57 (McGowan, 12/19).

Ultimately, Nachtsheim decided, over McGowan's objection, that the price of the Lachmar ships would not be reopened to *845 account for the cancellation of the Cherokee ships. There is no evidence in the record that Nachtsheim based his decision on any statement by McAndrew that QSD's labor hours had increased. Tr. 952-59 (McGowan, 12/19).

The Government asserts that QSD created an obligation to disclose the 1977 material purchase commitments by reason of a letter that H.M Class of QSD wrote to McGowan on November 29, 1976. Class' letter stated that "[a]t this time, General Dynamics ... does not anticipate ordering material until the award of construction differential subsidy by the Maritime Administration." JX257. Class was not involved in the negotiations with MarAd in any way. Moreover, there is no evidence in the record that the statement was false, and there is no evidence that, on November 29, QSD anticipated that the contracts would not be signed until July 1977 and intended at that time to purchase material before the contracts were executed. In addition, McGowan never asked to be told if QSD changed its mind. Tr. 965-66, 969-76 (McGowan, 12/19). Furthermore, the statement by Class had nothing to do with the price of the Lachmar ships and pertained to an unrelated issue of compliance with EEOC regulations GD214/JX779; Tr. 966-70 (McGowan, 12/19). Finally, the Government has failed to prove that McGowan, or anyone else at MarAd, relied on the Class letter in July 1977 when MarAd decided not to reopen the price of the Lachmar ships.

B. Sphere Labor Estimate

QSD's sphere production effort began in November 1972 when QSD contracted with PDM, its subcontractor for spheres, to provide three shipsets of spheres. The sphere production work proved to be much larger and more complex than PDM had anticipated. In June 1974, MarAd representative Norm Hammer reported that PDM "continually underestimated the magnitude and complexity of the sphere fabrication work." In October, 1974, MarAd representative Sam Wolkow observed that the situation at Charleston was "critical and continues to deteriorate and could jeopardize the entire LNG program." JX520; GD 363 at 10519. Faced with total disaster in December 1974 QSD terminated PDM's contract and stopped work. GD took over the Charleston Facility and decided to attempt to build the spheres itself.

Thereafter, QSD embarked on an extensive capital improvement program at Charleston, with an initial appropriation of $56.8 million. However, despite this infusion of capital, QSD also encountered substantial difficulties building the spheres, and repeatedly failed to meet its cost and schedule projections. For example, in November 1975, Veliotis told the Board of Directors that the first sphere would be completed in three months, by February 1976. On April 20, 1976, Arthur Andersen reported that QSD was then predicting that the first sphere would come out of the jig by the end of April 1976. Neither of these projections proved even remotely accurate. JX61 at 136093; JX106 at 002.

With the aid of a Swiss consulting firm named Vevey, QSD began the final welding effort in January 1976. However, Vevey's initial attempt, in March 1976, at final welding the first sphere soon proved disastrous. One month later, in April 1976, MarAd noted that although the first sphere was 50% welded, no welds had "been submitted for inspection nor accepted by the Owner or the regulatory agencies." In fact, none of the welds were of acceptable quality, and every inch of weld had to be chipped out. This was an extraordinarily painstaking, time consuming, and costly process. QSD had to reweld the sphere three times. GD363 at 10666; Tr. 3050-52 (Tovar, 2/11).

By the summer of 1976, the Charleston facility had reached another state of crisis. Its production welders had not completed the final welding of a single seam. In May 1976, MarAd knew that the problems at Charleston were so severe that QSD again was contemplating suspending work until it could review the welding situation. GD363 at 10674; JX698 at 12180-81; GD161. In a June 30, 1976 memorandum, written at the same time QSD prepared its proposal support, *846 Wolkow gave the welding effort a bleak assessment. See JX549.

On July 22, 1976, three days after QSD submitted the Proposal Support, MarAd Administrator Robert Blackwell, his deputy Howard Casey (who later chaired the Subsidy Board's review of the Lachmar CDS award) and Pross visited the Charleston facility and found there were still "no signs of improvement which would support any type of production flow." GD363 at 10704-05. Starting in the summer of 1976, QSD made a variety of changes at the Charleston facility to try to break the logjam. Rather than relying exclusively on Vevey, QSD looked to other divisions of General Dynamics for expertise in metal working. GD sent employees from the Convair and Ft. Worth Divisions to Charleston to try to improve welding techniques and to solve the production problems. GD179; GD182.

The GD team sent to Charleston identified numerous problems, some of which were easily solved, but some of which were intractable. Among the more difficult problems was the shortage of skilled welders. Not only did Charleston's welding force have little experience in welding aluminum, it also lacked supervisors skilled enough to teach the art. GD152; JX698 at 12195. In August 1976, the Government's regulatory authorities approved a change in the method QSD used to test the welds on the spheres, and in early September 1976, Wolkow reported that this change in the inspection procedures led to a dramatic improvement in the number of welds that were accepted by the owner's representative and the regulatory authorities. Wolkow also reported, however, that fourteen major production steps remained to be attempted for the first time, and that "it is anticipated that unexpected problems will develop requiring corrective action and/or revisions." GD363 at 10718-20.

Charleston finally completed the first sphere in December 1976. The sphere consumed over 200,000 manhours, more than twice the number originally estimated. GD262 at D010194; GD397 at 137631; GD398 at 134073. Welding problems did not end with the completion of the first sphere. In fact, it was not until late December 1976 or early January 1977, well after QSD and MarAd reached agreement on price, that QSD submitted a seam to the regulatory bodies for approval that did not require repair. JX543 at 14079.

Because of the amount of rework required on the first sphere, QSD discarded that sphere for purposes of applying a learning curve to estimate the sphere labor hours for Lachmar ships. JX180 at 27. QSD began its learning curve with the second sphere as a "theoretical first sphere" at 160,000 hours. Id. From this theoretical first sphere through the 20th sphere, QSD applied a 92% learning curve, which was a composite or weighted average of individual learning curves developed for each of the major areas of sphere construction work: fabrication and subassembly (96%); assembly and disassembly of jigs (86%); welding assembly (90%); hydrotest and ship-out (95%); dome (estimated by components). Id.; JX196/GD462 at 102898. GD leveled the hours from the 20th to the 60th sphere at the 20th sphere estimate. The resulting estimate for the 11th and 12th ships was 111,584 hours per sphere. JX180 at 27. Using estimating worksheets prepared by McAndrew, Murphy demonstrated convincingly and at length that each component of the sphere labor estimate was reasonably prepared. JX172; Tr. 3422-33 (fabrication and subassembly work); Tr. 3433-40 (assembly and disassembly of jigs); Tr. 3440-49 (welding assembly); Tr. 3449-52 (hydrostatic test and ship-out); Tr. 3453-57 (dome assembly). The Government did not challenge or rebut any of the specific estimates that McAndrew prepared, and the Government failed to show that any of his judgments were unreasonable or inconsistent with the existing factual data.

Moreover, QSD's estimate that the number of sphere hours would level off at the 20th sphere also was reasonable and free of false statements. Tr. 3785-87 (Murphy 2/19). Most importantly, the judgment that QSD made was fully disclosed to MarAd. JX196 at 102898

*847 In sum, there is abundant evidence that QSD's sphere labor estimate was prepared in an ordinary, and reasonable manner, and that it was consistent with the existing factual data that was reported in the Charleston facility's labor progressing system. Tr. 3458 (Murphy, 2/14).

We do not find Rylander's conclusion, that QSD's sphere labor estimate was fraudulent, persuasive. Rylander almost exclusively based his conclusion on the fact that QSD had a budget for sphere labor of 63,000 hours per ship. Tr. 1920 (Rylander, 1/10).

In part, Rylander engaged in a false comparison. As opposed to the QSD budget estimate, the proposal support included the dome assembly as part of sphere labor, even though the dome actually was constructed at Quincy. This item was not included in the budget. JX196 at 102898; JX183 at 104973; Tr. 3418-19 (Murphy, 2/14); Tr. 1843-44 (Rylander, 1/10). More importantly, there is no evidence to support the budget estimate as a benchmark for determining fraud as of the summer of 1976. The budget estimate can be traced back at least to June 1975, soon after QSD took over the Charleston facility and long before the problems and production delays of late 1975 and 1976 occurred. In other words the budget was not based upon data current as of the summer of 1976. On June 13, 1975, Veliotis told the Executive Committee of the Board of Directors that, while cost estimates of the spheres were predicated upon 100,000 manhours per sphere, he was "hopeful" that "through efficiencies" this figure could be reduced to 62,500 manhours per sphere. GD397 at 137631; GD398 at 134073.

Rylander acknowledged that he had no idea how the budget was prepared, that he could not demonstrate that there was a reasonable basis for the 63,000 hour budget figure, and that he had found no data, as of the time QSD submitted the proposal support, that a competent cost estimator could use to obtain the 63,000 hour forecast. Tr. 1912-14 (Rylander, 1/10). Moreover, Murphy confirmed that he found nothing to support the budget estimate that would remotely compare to the detailed factual information in JX172 that McAndrew used to prepare the proposal support.

Fritz Tovar, the General Manager of the Charleston facility, testified forcefully that it was his view at the time that it would take a "miracle" to reach the budget forecasts, which included a projection that QSD would build the fifth sphere for 75,000 hours. He was under substantial pressure from Veliotis however, to maintain those targets. Tr. 3072-73; 3053 (Tovar, 2/11). Moreover, as noted above, throughout 1975 and 1976 QSD routinely failed to meet virtually every production forecast for the spheres. Tr. 3073 (Tovar, 2/11).

The Government also suggests that even if the proposal support was not fraudulent when it was submitted in July 1976, it somehow became fraudulent as a result of progress that was made on the spheres during the autumn of 1976. The Government primarily attributes this progress to the enclosure of the main assembly hall, and claims that QSD alone was aware of the significance of this event.

First of all, MarAd was fully aware of every step of progress on the spheres and could have requested a reevaluation of the sphere labor estimate at any time. Pross and others at MarAd received progress reports from Charleston literally on a daily basis. MarAd was aware that the welding of the spheres had been adversely impacted by the fact that the main assembly hall was not enclosed prior to June 1976. JX550; JX544 at 9415; GD171; GD174; GD195.

Moreover, QSD provided information to MarAd that allowed the agency to quantify whatever improvements were occurring. On November 4, 1976, Wolkow reported: "Data developed by General Dynamics on the cost of welding spheres # 1 and # 2 show a cost reduction of 50% to complete the erection seam welding on sphere # 2.... Also, a cost of reduction of 16% was realized in fitting and tack welding of sphere # 2." GD363 at 10745.

Finally, the Government contends that the actual sphere labor totals prove that the estimate QSD gave MarAd was *848 unreasonable. However, the question before the Court is the reasonableness of QSD's sphere labor estimate in light of the data that existed in 1976 (which is recorded in detail in JX172), not the data that exists in 1992. GX391.[5]

In addition, there are at least two reasons to question the reliability of the purported actual sphere totals. First, the substantial decrease in labor hours at Charleston was almost completely offset by a corresponding increase in the direct and overhead labor rate. As a result, QSD's total sphere and labor costs were almost identical to that projected in the MarAd bid. Murphy testified that it would be necessary to look at the accounting records to make sure that costs were not reclassified from direct to indirect, which would distort the actual sphere labor hour totals. Unfortunately, however, the data necessary to conduct that analysis no longer exists. Tr. 3801-02 (Murphy, 2/19).

Second, the actual sphere labor hour totals set forth in GX391 fall on a 56% learning curve, beginning with the 11th sphere. In other words, even eliminating whatever problems might have existed on the first few spheres, the actual hours suggest an improvement trend in the middle of the program of 56%. Murphy testified that a learning curve of that magnitude is virtually unheard of. Murphy's testimony again raises questions concerning the comparability of the proposal support estimates with actual labor hour data, and the ability of QSD to predict in the summer of 1976 the actual sphere hour totals. Tr. 3804-05 (Murphy, 2/19). Therefore, we do not find the discrepancy between the sphere labor hours in the proposal support and the actual labor hours as persuasive evidence of fraud.

C. Operations Recurring Hours Estimate

The operations recurring hour estimate that QSD submitted to MarAd to support the reasonableness of the Lachmar price had two components: (a) an estimate that it would take 3.852 million recurring hours (rounded down to 3.8 million hours for the purposes of estimating the Lachmar ships) to build the first LNG ship; and (b) an estimate that the ship-to-ship labor hour reductions would fall on a 94% learning curve. The mathematical application of these two factors produced an estimate for the Lachmar ships that averaged 3.056 million recurring hours per ship. JX180 at 2, 15.

Each of the components of the labor hour estimate for the Lachmar ships also were estimates. The first ship still was far from complete (indeed, the first sphere would not be completed and delivered for almost six months), and with no ships finished, it was impossible to measure the actual learning between completed ships. Tr. 3256 (Murphy, 2/13). The labor hour aspect of the proposal support was purely "fixed price." To the extent that QSD underestimated the number of labor hours, it bore the sole risk of the resultant overrun. The escalation protection of the proposal applied only to the labor rate. Tr. 3243 (Murphy, 2/13).

The estimate that it would take 3.852 million recurring hours to build the first ship was reasonable, and has not been challenged by the Government. Tr. 3249-53 (Murphy, 2/13). Indeed, the available data supported even a higher estimate. Tr. 3250-53 (Murphy, 2/13). Thus, the evidence in this regard is inconsistent with the Government's basic claim that QSD sought to inflate its estimate of the first ship to defraud MarAd.

McAndrew prepared the cost estimating support for the 94% learning curve used to develop the recurring hour estimate for the Lachmar ships. He prepared a document, which appears within a group of papers marked at trial as GD27, that identifies and explains how he prepared that estimate. GD27 at XXXXXX-XX; JX720 at 358. (McAndrew dep.); Tr. 189-91 (Renn, 12/11). Pages 11004 through 11009 of GD27 show that McAndrew estimated a 94% overall learning curve by preparing separate estimates of learning curves for each of the *849 three-digit accounts into which QSD divided the work of building an LNG ship. The rationale for each estimate is identified on the document.

Murphy "reversed-engineered" GD27 to verify and further identify how McAndrew derived the learning curves for each three-digit account. Tr. 3260-63, 3283-88 (Murphy, 2/13). McAndrew sought to estimate the learning that was occurring between the first two ships. With respect to accounts constituting 71% of the ship, McAndrew chose as the learning curve the trend indicated in the OLAR (Operations Labor Analysis Reports) system from the first to the second ship, with appropriate adjustments. Tr. 3283-84 (Murphy, 2/13). For the remaining accounts, where there was little or no experience on even the first two ships (such as the cargo accounts), McAndrew estimated the learning that would be achieved in that account. Generally, he adopted the learning curve from a similar account. Tr. 3283-85 (Murphy, 2/13).

In preparing the estimates of learning for each three-digit account, McAndrew began with the estimates at completion that appeared in the OLAR system. As identified on GD27, McAndrew then made adjustments to those OLAR estimates. For the structural hull accounts, McAndrew adjusted the indicated learning trends shown in the OLAR system to account for phenomena known as "tail-up" and "pick-up." GD27 at XXXXXX-XX; GD446; Tr. 3263-68; 3282-83 (Murphy, 2/13).

"Tail-up" is a widely recognized phenomenon in which the number of hours needed to complete the last few percentage points of completion of a project is greater than the number of hours recorded for the earlier percentage points of completion. Tr. 3270-71 (Murphy, 2/13); Tr. 2868-69 (Beggs, 2/6); Tr. 1251-53 (Cowan, 1/6); Tr. 269-70 (Renn, 12/11); Tr. 352-53 (Renn, 12/12). "Pick-up" is a part of tail-up. Although the terms at times are interchanged, "pick-up" technically is the common phenomenon in which additional hours are charged to work packages after they have been reported to be 100% complete. Pick-up can occur for many reasons, such as a mistaken assessment that the work was complete when it was not, or because a subsequent inspection revealed defects in the work that required correction. Tr. 3269-70 (Murphy, 2/13).

The record demonstrates that QSD was experiencing both "tail-up", see, e.g., JX220 at 057; JX175 at 138841; GD447; Tr. 3272-75 (Murphy, 2/13), and "pick-up", see, e.g., JX220 at 041, in QSD's construction of the LNG ships. McAndrew specifically noted these phenomena in preparing his estimates of the learning that were going to be achieved on the three-digit accounts. See GD27 at 110005.

There is other support in the record for the necessity of making adjustments to the OLAR estimates at completion. Tr. 3537-41 (Murphy, 2/18). For work that had not been started or was less than either 15 or 25% complete (the record does not establish the exact percentage), the OLAR system used "budget" numbers, with at most minor adjustments for "target performance." By the summer of 1976, the unreliability of the budget was well established. Moreover, because of the substantial errors in the budget, the relative weighting of tasks in the OLAR system was inaccurate. JX220 at 014; Tr. 644, 654-57 (Whittemore, 12/17); Tr. 849 (Whittemore, 12/18); Tr. 3371-78 (Murphy, 2/13). The unreliability of the OLAR estimates is demonstrated by the fact that, even though the first and second ships were nearing completion as of September 1976, estimates for those ships based on the OLAR system still were materially understated as much as 400,000 hours for the second ship. Tr. 866-74 (Whittemore, 12/18).

General Dynamics graphically demonstrated on GD482 the limitations of the OLAR estimates at completion for the first five ships as of June 26, 1976. The unadjusted OLAR estimates suggest increasingly steep learning on the LNG program that begins to approach a vertical line between the third, fourth, and fifth ships. GD482; Tr. 4243-45 (Murphy, 2/25). Even for that day, the OLAR system was extremely simplistic, using only linear extrapolations and *850 unreliable budgets to calculate estimated costs at completion.

The Government witnesses, including Rylander, conceded that adjustments would have to be made to the OLAR numbers to produce a reliable estimate. Indeed, McGowan criticized an analysis that one of his subordinates had prepared which, like the OLAR system, merely divided actual hours incurred to date by reported progress as "a mechanical extension that gave no credence to anything except arithmetic." JX697; Tr. 1039-42 (McGowan, 12/19); Tr. 657-58, 663-64, 668 (Whittemore, 12/17); Tr. 871 (Whittemore, 12/18); Tr. 2250-51 (Rylander, 1/29). MarAd was aware that the estimates of learning McAndrew prepared included adjustments for phenomena like tail-up and pick-up from pages 110004 through 110009 of GD27. While the record is far from clear, the Government has not carried its burden of showing that McAndrew failed to give this document to MarAd. See JX720 at 348-49, 358 (McAndrew Dep.).

The specific adjustments made by McAndrew which again are indicated explicitly on pages 110004 though 110009 of GD27, produced an overall estimate of approximately 94% learning between the first and second ships. At trial, General Dynamics demonstrated that this estimate was extremely accurate. In fact, Rylander conceded in his rebuttal testimony that the judgments McAndrew made on the first two ships were reasonable. Tr. 3288-3300 (Murphy, 2/13); Tr. 4350-51 (Rylander, 2/25).

The Government maintains that GD's use of a two ship learning curve was unreasonable and that McAndrew also should have used data from the third hull in calculating the learning curve. While there was conflicting testimony from GD's and the Government's experts over the propriety of using a two ship learning curve, we find that the Government has not successfully carried its burden to show that GD's use of a two ship learning curve was unreasonable. It was legitimate on several grounds for McAndrew to use the estimate of learning that occurred between the first two ships as his estimate of learning for the remainder of the program. While McAndrew also could have used other estimating methods, his choice of methodology is not proof of any fraud. Tr. 3305-18 (Murphy, 2/13).

First, learning curve theory is predicated on the notion that the learning that occurs between the first two units will be repeated between the second and the fourth units, and again between the fourth and the eighth units, and for each subsequent doubling of quantity. Thus, McAndrew did nothing more than apply standard learning curve theory. Given the fact that McAndrew had very little hard data QSD had not completed a single ship, and estimates at completion for the ships under construction were constantly increasing it was reasonable for him to utilize widely accepted learning curve methodology. Tr. 3305, 3315-18 (Murphy, 2/13).

Second, it was reasonable for McAndrew to focus on the first two ships because virtually all of the hard data at his disposal was associated with those two ships. As of June 26, 1976, QSD had incurred a total of 8.177 million recurring hours on the LNG program. Over 70% of the total or 5.761 million recurring hours had been spent on the first two ships. Moreover, as of that date the third ship was reported in the OLAR system to be only 54.69% complete, and over 60% of the work on the third ship had been concentrated in only two accounts, the midbody and the stern. JX175 at 138840, 138843.

Third, QSD's prior experience showed that the learning that occurred between the first two ships was a reliable indicator of the learning that could be expected to occur on the program as a whole. In fact, prior to the LNG program, QSD had never been involved in a multi-ship program in which the learning achieved between the first two ships understated the learning that would be achieved on the remainder of the program. GX21 at 109644; GD454; GD455; GD456; Tr. 3306-14 (Murphy, 2/13).

Moreover, there is an additional reason why it was reasonable for McAndrew to *851 use the 94% learning that he forecast between the first two ships as his estimate for the remainder of the program. A 94 or 95% learning curve is standard for shipbuilding. Tr. 1532-33 (Rylander, 1/7); Tr. 2383 (Rylander, 1/30).

The Government contends that MarAd may not have understood that QSD used a learning curve to prepare QSD's estimates at completion for the third, fourth, and fifth hulls listed in JX96. However, the Government apparently concedes that if it had had pages 110004-110009 of GD27 it would have been able to ascertain that the costs at complete by account in JX196 "were in major part derived from a two ship learning curve." See Government's Proposed Post-Trial Findings of Fact at ¶ 190. As we have already concluded, the Government has not satisfied its burden of proof with regards to its asserted lack of possession of these pages. Therefore, we find that although the JX196 can be read as the Government suggests, GD27 pages 110004-110009 put the Government on notice that GD was using a two ship learning curve to estimate the cost at completes ("CAC's") of the third, fourth, and fifth ships. Moreover, whether or not MarAd understood how McAndrew prepared his aggregate estimate of 94% learning, the important point is that for all the reasons set forth above, the estimate was reasonable and consistent with both existing historical factual data, as well as generally accepted cost estimating practices.

As evidenced primarily by Gessow's notes and summaries, MarAd reviewed certain aspects of QSD's proposal support with tremendous care. MarAd frequently requested and received additional supporting data from the shipyard and other sources. See, e.g., JX609/GD402 at 0881, 0893, 0911; JX611/GD403 at 0865; GD22 at 1391-93 (wire estimate); JX609/GD402 at 0876, 0894, 0911; JX611/GD403 at 0860-61; JX615 (cable estimate); JX611/GD403 at 0866; JX609/GD402 at 0877-78, 0894, 0911 (paint estimate). There is no evidence that QSD ever denied MarAd any information MarAd requested. JX609/GD402.

Gessow's notes and summaries show that MarAd conducted very little analysis of QSD's recurring labor hour estimate, and the estimate was not even discussed during the negotiations. Instead, in Gessow's negotiation notes for November 18, 1976 the day on which the parties reached agreement on price there is the following entry: "JJM [John J. McGowan] note: we never raised issue as to the hours or overhead, we think that your estimate is reasonable w/r to that." JX609/GD402 at 0920. Indeed, on direct examination McGowan admitted that he did not use the learning curve as a basis for his judgment that the labor hour estimates for the intermediate hulls were fair and reasonable. Tr. 918, 920-21 (McGowan, 12/19).

It is understandable why McGowan did not question QSD's labor hour estimate. QSD projected that the number of recurring hours spent on each of the Lachmar ships would be 796,000 hours or more than 20% less than the current estimate for the still incomplete first ship. The only question raised by the Government is whether QSD should have projected even greater improvements in its future performance, not whether QSD improperly inflated existing costs. In a program designed to support the American shipbuilding industry, it is understandable why McGowan was satisfied with an estimate that was more than 20% below what the shipyard had ever achieved or was expected to achieve on the ship nearest completion.

For all these reasons, the Court finds that MarAd did not rely upon the method that QSD used to estimate the number of labor hours for the Lachmar ships, or any estimates that QSD submitted for the intermediate ships under construction.

QSD's labor hour estimate for the Lachmar ships was based on two estimating parameters: An estimate for the first ship of 3.8 million hours, and a 94% learning curve. As noted above, the Government does not challenge the reasonableness of the 3.8 million hour estimate for the first ship, and its expert volunteered that a 94% learning curve is the "industry standard." Moreover, it is a fact that the first ship *852 actually required more than 3.8 million hours, and that the first two ships actually fell on a 93% learning curve. Therefore, the Government's entire theory of fraud appears to be necessarily predicated on its contention that QSD knew in the summer of 1976 that the labor performance of the first two ships was wholly aberrational and would not continue on the remaining ships. JX180 at 2, 15; GD266 at 122449; Tr. 2243-45 (Rylander, 1/29); Tr. 1532-33 (Rylander, 1/7); Tr. 2383 (Rylander, 1/30).

Contrary to the Government's contentions the record clearly establishes that the shipyard had an entrenched history of chronic labor problems, Tr. 2833-36, 2847-51, 2896-92 (Beggs, 2/6), and that the labor build-up in the summer of 1976, when QSD prepared its price support for the Lachmar ships, was far from over. See JX180 at 26; JX308 at 106; GD363 at 700052; Tr. 3232-33 (Murphy, 2/13).

Moreover, Rylander, the Government's expert, conceded away the Government's theory of liability when he agreed with GD's assertion that "the whole question of hiring, training, and assimilation of tradesmen was a continuing problem throughout this period of time, well after General Dynamics and the Maritime Administration reached agreement with respect to price." Tr. 2268 (Rylander 1/29); see 1832 (Rylander, 1/9); Tr. 2256 (Rylander, 1/29); Tr. 2344 (Rylander, 1/30).

Decisively, Murphy, GD's expert, convincingly testified that the first two ships were not aberrational. Tr. 3318-19 (Murphy, 2/13); Tr. 3859 (Murphy, 2/19); Tr. 2949-50 (Beggs, 2/6).

Finally, we do not find convincing Rylander's parametric analysis which purports to show that a reasonable estimate of recurring manhours for the structural accounts of the LNG ships was between 35 and 55 manhours per ton. Rylander's analysis is based on a comparison of an LNG ship to the "Lykes" ships that QSD built for the Lykes Steamship Company, and the analysis assumes that (1) the hull of an LNG ship was less complex than that of a Lykes vessel; and (2) that, apart from the spheres and cryogenics, an LNG vessel was no more difficult to construct than an ordinary tanker. However, we find that both of these assumptions are incorrect.

First of all Rylander knew virtually nothing about either the Lykes or the LNG ships. He had never seen one of the Lykes ships, had never walked a Lykes ship, and never reviewed its specifications. Similarly, Rylander never walked an LNG ship and did not even bother to look at the ship's design drawings. Tr. 1757, 1799 (Rylander, 1/9); Tr. 2294-95 (Rylander, 1/29); Tr. 2581 (Fisher, 2/4).

Moreover, one of GD's experts, Fisher, testified that LNG ships were considerably more complex than the Lykes ships, Tr. 2581-85 (Fisher, 2/4), and MarAd's own experienced cost estimators, Schimler and McGowan, concluded in 1972 that "[t]he internals of the LNG design structure are more complicated than a standard tanker." JX515 at 0559.

For these reasons, the record does not support the Government's assertion that QSD should have estimated that the structural work on an LNG ship could be completed for between 35 and 55 manhours/ton.

Although QSD's internal operating plans projected that the 11th and 12th LNG ships would be built in 2.5 million hours, nowhere does there appear any estimating detail to support this 2.5 million hour figure. The number first appears in an October 1, 1974 estimate. At that time, the first hull was only 23% complete and most of the significant problems that were going to arise on the LNG program had not yet occurred. JX747 at 8384; GD363 at 10526; Tr. 3459-60 (Murphy, 2/14). Substantial evidence in the record establishes the unreliability of the labor hour estimates in QSD's operating plans.

First of all, the evolution of the operating plan estimates over time reflects a "fan-tail" effect. Even though QSD management was repeatedly forced to increase its estimates in the current period for the ships under construction, it refused to make any significant modification to the estimates that existed for future periods. Veliotis refused specific requests by *853 Grimes to add hours to the operating plan. GD467; JX259; Tr. 3532-38 (Murphy, 2/18).

Moreover, as the operating plan estimates changed over time, they moved further and further away from what would be expected by a normal learning curve progression. Arthur Andersen recognized that the division's plans were based on operational goals, not learning curve theory. GD467; JX775 at 93; Tr. 3535-36 (Murphy, 2/18).

The unreliability of QSD's operating plan estimates was verified by Philip Cowen, a former corporate executive of General Dynamics who was called as a witness by the Government. Cowen explained that QSD had a history of providing forecasts that were overly optimistic relative to the performance that the shipyard would achieve. Moreover, Cowen testified that the shipyard usually "had little hard evidence. It was a degree of optimism that [I] generally found unwarranted by the circumstances." Tr. 1306-08 (Cowen, 1/6).

It is true that QSD expressed its operating plan position in representation letters to Arthur Andersen. On the basis of the aggressive assumptions in the operating plan, including the 2.5 million hour labor estimate, QSD projected substantial profits on LNG ships nine through twelve. Tr. 1909-10 (Rylander, 1/10); Tr. 4075 (Murphy, 2/21). Arthur Andersen was aware that the estimates it received in the representation letters were the same as the shipyard's operating plans. Tr. 572-73 (Whittemore, 12/13); Tr. 4075 (Murphy, 2/21).

What is significant, however, is that General Dynamics was unwilling to book the profit that QSD forecast in its operating plans and its representation letters to Arthur Andersen. Instead, throughout 1976 and the first half of 1977, GD recorded no profit on the LNG program and continued to write off period costs. GD did so despite the fact that it would have been permitted to record a profit under standard accounting rules had it "believed" QSD's operating plan profit estimates. Tr. 682 (Whittemore, 12/17). John Whittemore, a former auditor with Arthur Andersen who worked on the account and whom the Government called as a witness, described GD's unwillingness to record the profit that QSD forecast as unusual. Tr. 589 (Whittemore, 12/13); Tr. 683 (Whittemore, 12/17).

In effect, what GD did was treat the projected profit for book purposes as zero, thus effectively creating a reserve for the full amount of the profit forecast by QSD. Tr. 717-18 (Whittemore, 12/17). Thus, GD acted in a way much more conservative or pessimistic than that suggested by QSD's representation letters to Arthur Andersen. In these circumstances, the Court finds that there is no significance to the representation letters in terms of the reliability of QSD's forecast. As Whittemore testified: "It really doesn't make a whole hill of beans if they're projecting 30,000,000 of profit on an 8 ships program or 200,000,000 on a 12 ship program if they're picking up no profit on the books and records of the accounts. A hundred percent of zero is still zero." Tr. 718, 755-56 (Whittemore, 12/17).

QSD's operating plan forecasts also were expressed in various memoranda to General Dynamics' Board of Directors, in connection with the Board's consideration of whether or not the company should invest in the Lachmar partnership. Although the memoranda and financial analyses were prepared by Cowen or other corporate officers, Cowen testified that he simply included the Division's construction profit forecasts and did not do any independent analysis of them. JX71; JX136; JX339; JX352; Tr. 1328-29 (Cowen, 1/6).

The operating plan forecasts are not more reliable because they were communicated to the Board. Beggs, a GD board member during the events that occurred in this case, testified that he did not consider the construction profit forecasts in the memoranda to be "reliable at all." He added: "Nor do I think ... anybody else did." Tr. 2845 (Beggs, 2/6). Beggs explained:

The yard had never met any of its forecasts from the time we bought it in '64 until this period of time. To be sure, we *854 had installed new management. Taki [Veliotis] came in '73. Taki made forecasts and he didn't meet any of his forecasts either. The yard in short was finding it extremely difficult in this time period to meet any of the forecasts of business work that they had in the yard. Budget projections would be exceeded, schedule dates would be missed and, as I said earlier, we had a continuing situation of labor militancy at Quincy which made estimates difficult to achieve.... The situation at Charleston at this time was very uncertain at best. We were having continuing welding problems with the spheres.... All of those problems were present at this time, and while a great deal of optimism was being expressed by the yard and by Veliotis that, you know we would eventually solve these problems, we didn't know on what schedule.

Tr. 2845-47, 2887, 2898-99, 2923-24 (Beggs, 2/6).

Assuming that the Board did not believe that GD was going to earn construction profits, there were several other reasons that supported the board's decision to go forward with the project. First, the company had made a substantial investment both at Quincy and at Charleston, and needed additional work just to keep the shipyard open. In addition, as a result of the energy crisis that persisted at that time, the potential demand for LNG ships in the future was great. As a result, there was value in keeping the line running and the shipyard open. The Board also took into account the sizable profit that was projected from the operation of the Lachmar ships and the expectation that some profit would be earned building the ships. JX71; JX136; JX339; JX352; Tr. 2839-41; 2843-44, 2848-50 (Beggs, 2/6).

While the record is far from clear, we find that the Government has failed to prove by a preponderance of the evidence that the GD board relied on and adopted the forecasted Lachmar construction profit as the reason to go forward with the construction of the Lachmar ships.

Arthur Andersen had its own "position" regarding labor hour estimates for the first eight LNG ships. Its position generally was similar to that of QSD for the first ships that had substantial progress; for the remaining ships, however, Arthur Andersen, unlike QSD, adhered to a strict learning curve. With this one significant change, Arthur Andersen's position was not materially different from QSD's operating plan position. GD384; JX775; Tr. 665-66, 744-45 (Whittemore, 12/17); Tr. 865 (Whittemore, 12/18).

Arthur Andersen's estimates were based on a very limited review of QSD's operations. See Tr. 536-37 (Whittemore, 12/13); Tr. 839 (Whittemore, 12/18); Tr. 810 (Whittemore, 12/18). It did not conduct its own, more rigorous review of QSD's estimates because GD was not reporting any profit on the basis of those estimates and was continuing to write off period costs. Tr. 717 (Whittemore, 12/17); Tr. 839 (Whittemore, 12/18).

In conducting its limited review, Arthur Andersen essentially accepted the OLAR system and utilized it as a basis for its position, and did not prepare its own "bottoms-up" estimate despite the known frailties in the OLAR progressing system. Whittemore confirmed that, in essence, what Arthur Andersen did was check QSD's position which, as noted above, was based on illusory operational goals and not strict learning curve theory against the position that was suggested by QSD's own progressing system. Tr. 513, 554, 579 (Whittemore, 12/17); Tr. 861, 865 (Whittemore, 12/18). Arthur Andersen recognized that tremendous risks remained in its labor hour estimates and Whittemore emphasized that "you are dealing with a program with tremendous amounts of risk factors and you are dealing with estimates." It concluded that the risks were such that it was entirely appropriate for General Dynamics not to record any profit in 1976 and the first half of 1977. Tr. 635-37, 682-84, 751 (Whittemore, 12/17). In addition, Whittemore confirmed that the type of information on which Arthur Andersen's labor hour position was based was not sufficiently reliable to use for the purpose of pricing *855 a new ship. Tr. 657-68, 663-64, 667-68 (Whittemore, 12/17); Tr. 808-09, 871 (Whittemore, 12/18).

Based on these facts the Court finds that there is nothing in Arthur Andersen's labor hour position that establishes that the different estimate in the proposal support was fraudulent.

The Government apparently contends that two learning curve analyses conducted by Grimes were QSD's "real" learning curves. Grimes prepared the first learning curve analysis in June of 1976 and concluded that QSD's "applied" learning curve on the LNG program was 86.5%. Grimes prepared the second learning curve analysis in November of 1976 and concluded that QSD's "actual" learning curve was 86.0%. JX183 at 104974; JX259 at XXXXXX-XX.

Rylander relied heavily on both documents to support the Government's notion that Grimes analyses were GD's "real" learning curves. However, Rylander did no analysis of the documents, and he had no idea how Grimes prepared them. As Rylander testified: "I don't know how those numbers were calculated, so I can only you know, I would have to speculate on how it was done." Tr. 1823-24 (Rylander, 1/9).

Murphy, on the other hand, reviewed the June 1976 learning curve, reverse-engineered it, and convincingly demonstrated that it is seriously flawed and cannot be relied on by this Court to support a finding of fraud. Tr. 3329-3363 (Murphy, 2/13). Moreover, although Murphy was not able to conduct the same review of Grimes November 1976 analysis because the relevant OLAR data was no longer available, Murphy was able to establish that the November 1976 analysis also cannot be relied upon to support a finding of fraud. GD 460; Tr. 3331-32, 3356-58, 3358-63 (Murphy, 2/13).[6]

The Government contends that GD was obligated to turn over to MarAd a labor hour estimate for the first eight ships that GD captioned "real world," and that apparently used an 87% learning curve. The Government maintains that GD's failure to do so constitutes fraud.

First of all, there exists no estimating detail or support behind the "real world" estimate that would establish that the "real world" estimate was any more reasonable than the operating plan, Grimes' estimates, or the estimate GD provided MarAd in the proposal support. Indeed, the "real world" estimates for the first eight ships are only slightly different from estimates generated by Grimes, at approximately the same time, in JX183. JX183 at 104974; JX152 at 115236.

The Government seems to argue that the "real world" estimate was the estimate that GD truly believed was correct, and that the estimate GD gave to the Government was not GD's "best estimate." While there is some evidence in the record to support the Government's contention that the term "real world" refers to what actually was happening in the shipyard, none of the Government's witnesses testified that they had any personal knowledge as to what the phrase "best estimate CAC real world" specifically meant in JX133 and JX171 at 109720. Indeed, McAndrew testified at his deposition that he did not know what was the intended meaning of this phrase. JX720 at 365 (McAndrew Dep.). This sort of sketchy testimony is not evidence we find sufficiently probative to support a finding of fraud by a preponderance *856 of the evidence. Therefore, we do not find that GD's failure to turn over the "real world" estimate to MarAd was fraudulent.

The Government further contends that the statement of reasons that QSD gave MarAd in the proposal support (JX180 at 16-25) for the increase in hours on the first ship was incomplete and misleading. Specifically, the Government asserts that QSD failed to acknowledge that the increase in hours was due to poor trade performance. As set forth above, however, the increase in hours resulted from other sources rather than from atypical trade performance. The proposal support accurately reported that the current estimate reflected the "actual scope of the work experienced on the first hull." JX180 at 19; Tr. 3834-69 (Murphy, 2/19).

In addition, the Government has not established any reliance or harm based on the reconciliation of hours. First of all, MarAd did not request the information. Moreover, the information QSD provided was obviously a gross, high-level summary with little detail. Tr. 3224-25 (Murphy, 2/13). It should have been apparent to MarAd that the specific items listed on JX180 for the structural account could not account for an increase of 900,000 hours. Tr. 3836, 3839 (Murphy, 2/19). Even Rylander admitted that apart from any disclosure by QSD, MarAd was well aware that the shipyard was attempting to rebuild its workforce and had experienced labor problems. Tr. 1544 (Rylander, 1/7). The Government did not even attempt to elicit from McGowan that he was misled in any way by the statement of reasons for the increase in hours.

The Government also contends that an inference of fraud may be drawn from the fact that, if QSD truly had believed it was going to experience a 94% learning curve on the LNG program, GD would have been required to book an additional $50 million loss in 1976. See Government Proposed Findings of Fact at 134. We are troubled by this seeming inconsistency. However, as GD argues, losses need not be recorded until they are known or reasonably verifiable. Given the uncertainty and risk that existed in 1976, the Government has failed to prove by a preponderance of the evidence that it was unreasonable for QSD to use an estimate for pricing purposes that was consistent with the current poor experience on the program. This is true even though the risk that the poor performance would continue was not certain or quantifiable enough to require the immediate recording of loss. Moreover, the accuracy of General Dynamics' financial statements is not at issue in this case. Tr. 1361-62 (Cowen, 1/6).

We observe that corporate policy with respect to booking large corporate profits or losses is a highly ad hoc decision, and we note that the Government did not sufficiently explore its experts' rationale for the experts' assertion that GD should have booked a $50 million loss if it truly believed that it was experiencing a 94% learning curve. The Government's failure to press the matter leaves us unable to determine under what specific circumstances booking such a $50 million loss would have been prudent. Although GD did not provide any rebuttal testimony, we are disinclined to find liability principally on GD's alleged failure to book such a loss.

The Government also relies on an isolated comment in JX152 referring to the 94% learning curve as a contingency. As noted above, both projected and actual experience on the first ships which supplied over 70% of the available data fell on or very near a 94% learning curve. Thus, the 94% learning curve was consistent with actual experience and was not a contingency in the ordinary sense of the term. JX152 at 115250; Tr. 3683-84 (Murphy, 2/18).

The Government also contends that the "Structural Hull Improvement Charts" that QSD submitted to the Board of Directors showed learning greater than 94%. Murphy established that these charts were unreliable for several reasons. Most significantly, they did not compare ships at the same point of physical progress so that relatively complex work on one ship was compared to relatively easy work on another. If the analysis was weighted properly, the improvement between the first and the *857 second ship reported to the Board in July 1976 would have become very similar to that reflected in the June 1976 unadjusted OLAR. GD461; Tr. 3363-78 (Murphy, 2/13).

D. The Operations Non-Recurring Estimate

QSD's proposal support included an estimate of (or a request for) 200,000 nonrecurring operations hours per ship. JX180 at 15. This estimate was uniquely a matter of judgment because the need for nonrecurring hours depended upon future events about which no one could make any prediction with reasonable certainty. These events included the length of the LNG program and the useful life of QSD's jigs and fixtures. Tr. 3391-92 (Murphy, 2/14).

From the very beginning, MarAd challenged the need for so many nonrecurring hours on the Lachmar ships and requested "a detailed explanation for [that] required level of effort...." JX182 at 0954. In response, QSD furnished a detailed breakdown of its estimated nonrecurring costs by three digit work account. JX196 at XXXXXX-XX.

McGowan recognized the need for some nonrecurring hours, but he questioned the number of hours, when they would be incurred, and how they should be distributed across the continuing program. JX609/GD402 at 0885-87; Tr. 3394-95 (Murphy, 2/14); Tr. 1133-1135 (McGowan, 12/20). The subject of nonrecurring hours surfaced at a number of different negotiation sessions and was one of the most vigorously debated topics during the negotiations. JX609/GD402 at 0885-87, 0918; JX245. On November 17, 1976, when QSD agreed to reduce the number of hours estimated for various three digit work accounts by 20%, the subject of nonrecurring hours remained open. The following morning, QSD agreed to a further reduction in nonrecurring hours to 121,000 per ship.

According to the Government's second expert, Flesher, nonrecurring hours are part of "most normal manufacturing processes." Tr. 4496 (Flesher, 2/27). Moreover, according to Murphy, if jigs and fixtures used in shipbuilding are not replaced, they eventually deteriorate to the point where metal being processed shows defects and loses tolerance. Tr. 3390, (Murphy, 2/14).

The evidence that QSD contemplated an ongoing construction program extending far beyond the Lachmar ships is both clear and uncontradicted. JX50 at 114596; JX55 at 114582; JX176; GD01; Tr. 2909 (Beggs, 2/6). Because the need for nonrecurring hours depended on future events that no one could predict the length of the LNG program, the rate at which the special purpose jigs and fixtures would deteriorate, and the point at which replacing the jigs and fixtures became more economical than repairing them much of the nonrecurring component of QSD's labor hour estimate was a contingency. Tr. 3388-94 (Murphy, 2/14).

QSD's estimate for nonrecurring hours represented a judgment that was reasonable in light of QSD's nonrecurring experience on the early ships. Tr. 3392-97 (Murphy, 2/14). In defending its estimate to MarAd, QSD disclosed its judgments and made no false statements. Tr. 3397 (Murphy, 2/14).

Rylander claimed QSD's estimate of nonrecurring hours per ship was "false and fraudulent" because QSD did not include any nonrecurring hours in the operating plan cost estimates for ships 9-12 that were incorporated in QSD's representation letters to Arthur Andersen. Tr. 2011-12, 2027 (Rylander, 1/28). Rylander admitted, however, that QSD never included nonrecurring hours in its operating plans and separately accounted for nonrecurring hours for the first eight ships by collecting them in the "Hull 40" account. Tr. 2041-42 (Rylander, 1/28). Moreover, there is no evidence in the record that Rylander did an independent analysis of the reasonableness of QSD's estimate for nonrecurring hours. Tr. 2040-43 (Rylander, 1/28). Rylander simply compared the number of nonrecurring hours he found in the operating plan to the number of nonrecurring hours he found in GD's submission to MarAd, saw that they were different, and concluded *858 from that difference that GD had committed fraud.

Accordingly, we find that the estimate for nonrecurring hours that QSD gave to MarAd was fair and reasonable.

E. The Engineering Hours Estimate

QSD's proposal support included an estimate of 47,000 engineering recurring hours for each of the Lachmar ships, and 75,000 engineering nonrecurring hours for the first one. JX180 at 15. From the very beginning, MarAd challenged the need for so many engineering hours for the 11th and 12th ships. In its request for additional information on July 22, 1976, MarAd asked for a detailed "breakdown" for the engineering effort described in QSD's proposal support. JX182 at 0954.

In response to MarAd's request, QSD furnished a detailed breakdown of its engineering hour estimate, by three-digit account. The submission identified the three digit work accounts that were expected to consume the greatest portion of the nonrecurring hours for the 11th ship. The accompanying comments made clear that these engineering hours were expected because, inter alia, time would be needed to educate Lachmar, the new owner of the ship, about the detail design characteristics of the LNG ship.

McGowan recognized the need for some engineering effort as a result of a new owner, but questioned the total in QSD's proposal support. Though McGowan understood the contingent nature of QSD's estimate, McGowan told QSD on October 21, 1976, that QSD would have to justify an engineering hour estimate higher than 25,000 to 35,000 per ship. JX609/GD402 at 0888. In analyzing QSD's request, McGowan observed that 170,000 engineering hours were equivalent to 85 engineering man-years. As QSD noted, however, this was equivalent to only 20-25 engineers working over a period of four years. Id.

On November 2, 1976, McGowan met with McAndrew to negotiate the items that remained open after the first round of negotiations. With respect to the engineering estimate, McGowan wrote that for the following ships, "about 40-45,000 engineering hours are required." McGowan further noted, however, that "General Dynamics maintains that since no other work other than LNG's is under construction, the entire engineering department must be charged to that construction. Furthermore, they must maintain a capability to process change, accomplish the alterations, etc. This argument has merit and we probably must back off from the repeat ship concept." JX240 at 1029.

On approximately November 11, 1976, QSD prepared a summary of the parties' remaining disagreements in anticipation of the resumed negotiations. In this summary QSD noted: "As the state of the art advances and as operational data is returned from the fleet impact on engineering will continue to be substantial." The summary, which QSD apparently gave to MarAd during the face-to-face negotiations on November 11, also noted: "The new owners have already indicated interest in a variety of changes; while many changes are reimbursable they are seldom phased to take perfect economic advantage of the on-roll manpower. In addition, certain irreducible levels of engineers must be available to respond to total program requirements of a critical but unscheduled nature." Attached to the summary was an engineering manpower profile showing QSD's "proposed engineering man hours leveled over a three year period to be equivalent to some 28 engineers." JX245.

During the negotiations on November 11, McGowan rejected QSD's request for a level of engineering effort equivalent to 28 engineers for three years. McGowan reiterated his position that the repeat ships should consume no more than 35,000-45,000 engineering hours in total. McGowan added that, even assuming a new owner, the total should be no greater than 65,000 per ship. JX609/GD402 at 0916.

On November 17, QSD repeated its argument that the nonrecurring engineering hours in its proposal support would be necessary because of new owner requirements. Citing his prior experience with Bethlehem Steel and Nassco, two yards *859 that had built ships for new customers, McGowan repeated his view that QSD's engineering estimate was excessive. JX609/GD402 at 0918. The next day, the parties compromised their differences as part of a global agreement on price. The final negotiated base price included 76,000 engineering hours per ship. JX609/GD402 at 0920; JX253; Tr. 3407 (Murphy, 2/14).

Murphy analyzed QSD's submissions in support of the request for 169,000 engineering manhours and found them free of false statements. Tr. 3408-09 (Murphy, 2/14). Murphy further testified that JX130 and JX130A, the "Engineering-Design Workload" charts, confirm rather than contradict QSD's submittal to MarAd. Although Murphy could not confirm the actual hours reflected in the charts, and believed them to be in error, the phasing was fully consistent with the phasing charts on page five of QSD's proposal support. Tr. 3412-15 (Murphy, 2/14).

Rylander claimed that JX130A proves that QSD misrepresented its estimated future engineering manhour needs to MarAd. However, Rylander never testified that QSD's engineering hours estimate was inherently unreasonable. Moreover, Rylander conceded that he based his finding of fraud solely on what he perceived as an inconsistency between JX130A and the engineering manpower profile in JX245. Tr. 2057-58 (Rylander, 1/28).

We find that there was no inconsistency between these two documents. Tr. 3413-15 (Murphy, 2/14). Moreover, Rylander conceded that QSD's request for nonrecurring engineering hours was clearly and separately identified, supported by argument, negotiated at length, challenged by McGowan on the basis of his experience with other yards, and ultimately compromised. Tr. 2064-65 (Rylander, 1/28). Additionally, Rylander also conceded that he performed no analysis of QSD's estimate of engineering hours and that his "reasonable range" of 18,000 to 47,000 hours per ship was based exclusively upon QSD's operating plan, Tr. 2068-70 (Rylander, 1/28), which, as we have indicated, was unreliable.

F. The Sphere Insulation and Joiner Estimates

During the 1970's, Frigitemp Corporation ("Frigitemp"), a New York corporation, was a subcontractor on GD shipbuilding projects at QSD and the Electric Boat Division. From 1974 to 1977, QSD awarded five subcontracts to Frigitemp: (1) in August 1974, a $4,669,448 subcontract for joiner work on LNG ships 1-8; (2) in January 1975, a $10,206,000 subcontract to insulate the spheres on LNG ships 1-3; (3) in June 1975, a $15,410,000 subcontract to insulate the spheres on LNG ships 4-8; (4) in October 1977, a $2,234,047 subcontract for joiner work on the Lachmar ships; and (5) in October 1977, $10,206,000 subcontract to insulate spheres on Lachmar ships. Veliotis was the person at QSD primarily responsible for awarding subcontracts for the LNG ships. GD Stip. Fact No. Supp. 1; J. 466 ¶¶ 6 & 7.

Gerald Lee ("Lee") was Frigitemp's Chairman and Chief Executive Officer and Mervyn Silver ("Silver") was Frigitemp's President. The Marine Division of Frigitemp ("FMD") was a subcontractor on major commercial and naval shipbuilding projects, and George Davis ("Davis") was a Vice-President of FMD. GX341 ¶ 7; GX368 at 4.

From 1973-1977, Veliotis had discussions with officers and employees of Frigitemp, including Davis, concerning the award of subcontracts for performing joiner and sphere insulation work on the LNG ships. Veliotis agreed with Davis and the others to award the subcontracts to Frigitemp in exchange for bribes or kickbacks to be paid to Veliotis and Gilliland, another senior official of QSD. JX466 ¶ 7. Pursuant to this kickback scheme, during the period of August 1974 to October 1977, Veliotis approved or caused the approval of the issuance of purchase orders to Frigitemp for subcontract work on all 10 LNG ships built by QSD, including the subcontract work for the Lachmar ships. In return Frigitemp and its former officers, including Davis, paid kickbacks to Veliotis and Gilliland. Moreover, in return for additional *860 kickbacks paid to Veliotis and Gilliland, and in order to prevent detection of the kickback scheme, Veliotis terminated all of QSD's subcontracts with Frigitemp in 1978 and awarded the subcontracts to Davis' newly formed company, Intersystem Design and Technology Corporation ("IDT") in 1978.

As discussed below in our conclusions of law, GD argues that it is not legally liable for the damages the Government suffered as a result of the Frigitemp kickbacks. However, assuming that GD is liable for the damages the Government suffered as a result of the Frigitemp kickbacks, GD and the Government dispute the amount of the damages. We find, for the record only, the facts as follows on this damage question.

The Government's damages for the sphere insulation and the joiner work will be the difference between GD's projected revenue for the Frigitemp subcontracts, and GD's cost at delivery estimates of the Frigitemp subcontracts. The greater the projected revenue and the lower the estimated cost, the greater are the Government's damages. We find that the projected revenue for the sphere insulation and the joiner work is $5,469,000 and $1,002,800 respectively. See GX128-F at 14 & Tr. 2194-97 (Rylander, 1/28) (additional escalation amounts listed in GX128-F at 14 should not be included in the calculation of the actual escalated costs of the sphere insulation and joiner work).[7] We find that GD's cost-at-delivery estimate for sphere insulation to be $2,934,000 and reject GD's contention that GD's internal sphere insulation estimate was a base price subject to escalation. See Tr. 2307-09 (Rylander, 1/30); Tr. 4377-78 (Rylander, 2/25). We also find that the GD's cost-at-delivery estimate for the joiner work to have been $894,000. Tr. 2210 (Rylander, 1/29). Therefore, the resulting damages are $2,535,000 per ship for insulation and $108,800 per ship for joiner work.

VII. Estimate Summary

Whether analyzed individually or as a totality, all of the estimates that QSD provided to MarAd, except for QSD's estimates for the Frigitemp joiner and sphere insulation work, were objectively fair and reasonable. This is the conclusion that Gessow came to in a 1985 analysis that he did for MarAd after Veliotis asserted to the F.B.I. that GD defrauded MarAd in connection with the CDS subsidy awards. See GD298 at 1, 2, 7. Indeed, one of GD's experts stated at trial that "the Lachmar ship prices were definitely fair and reasonable vis-a-vis the marketplace, and in fact were probably on the lower end of the market spectrum." Tr. 2549 (Fisher, 2/4); see also Tr. 2548-71 (Fisher, 2/4) (stating that the estimate that QSD gave to MarAd was fair and reasonable market price).

Based on the entire record in this proceeding, we find that the estimate QSD submitted to MarAd to support the negotiated Lachmar price was objectively reasonable and fair, and consistent with existing experience at the shipyard. Moreover, with the exception of GD's estimates for Frigitemp's joiner and sphere insulation work, we find that the Government has failed to prove by a preponderance of the evidence that GD did not believe that the estimates it submitted to MarAd as part of GD's proposal support were GD's "best" estimates. Finally, we find that QSD truthfully reported all the underlying factual data the actual hours incurred, the vendor quotations, the status of production at Charleston, and all other construction and finance matters on which GD's estimates were based.

Conclusions of Law

The primary issue in this case is whether the proposal support that GD provided to MarAd, as part of GD's application for CDS's, satisfied the reporting requirements of the Merchant Marine Act of 1936, as *861 amended in 1970 ("the act"). Because there are no cases that address the legal requirements governing information that shipyards must provide to MarAd when shipyards apply for CDS's, the instant action appears to be a case of first impression.[8]

I. GD Satisfied the Merchant Marine Act's Proposal Support Reporting Requirements

As noted above, the act requires a shipyard that has agreed with a ship purchaser on a ship price to "submit [to MarAd] backup cost details and evidence that the negotiated price is fair and reasonable." 46 App. U.S.C. § 1152(a) (ii). The words "fair" and "reasonable" indicate that the Secretary should use an objective standard when determining whether the negotiated ship price is acceptable, since to do otherwise would relegate each inquiry under the statute to a web of subjectivity and destroy the continuity and predictability of administration and enforcement. Similarly, these words also indicate that the cost estimates that a shipyard submits to MarAd must also be objectively reasonable. See generally United States v. Davis, 666 F. Supp. 641, 643 (S.D.N.Y.1987) (Weinfeld, J.) (holding that "MarAd subsidy contracts must be consistent with the enabling statute which authorizes their execution," and that therefore cost estimates that shipyards provide to MarAd must be fair and reasonable).

In the legislative history of the 1970 amendments to the act, Congress recognized that estimation of costs "is not an exact science and that reasonable men might differ over the estimate[s] reached." S.Rep. No. 1080, 91st Cong., 2d Sess., reprinted in 1970 U.S.Code Cong. & Admin.News 4188, 4201; see Tr. 3517-18 (Murphy, 2/18); Tr. 1755-56 (Rylander, 1/9). Thus, Congress was aware that for each ship construction project there exists a range of cost estimates that is objectively justifiable. Moreover, Congress was also aware that while cost estimators may agree on the breadth of this range, different cost estimators can come to different conclusions as to what they subjectively believe the actual cost of the project will be. For example, cost estimator A and cost estimator B may agree that ship construction costs on a project could range anywhere from $10 to $15 million, and that any cost estimate within that range is objectively reasonable. However, cost estimator A may subjectively believe that the project will cost $11 million, while cost estimator B may subjectively believe that the project will cost $14 million. If cost estimator A represents to MarAd that the ship construction project is going to cost $12 million, his estimate is objectively reasonable while being subjectively incorrect.

The Government appears to maintain that besides being objectively reasonable, the cost estimates an applicant provides to MarAd as part of the applicant's CDS proposal support must be the applicant's subjective best estimates of the actual costs of constructing the ship. We disagree. We believe that the structure of the CDS application process demonstrates that the act only requires a shipyard to submit to Marad objectively reasonable cost estimates that justify the ship price agreed upon by the shipyard and the ship purchaser through negotiation.

That the act apparently does not require a shipyard to submit to MarAd the shipyard's subjective best estimates of the actual costs of constructing the ship becomes clear when one considers that the act does not require that a shipyard and a ship purchaser calculate their mutually agreed upon ship price based on actual cost factors. Indeed, the Act implicitly envisions that the ship price may be market based rather than cost based. Had the act intended the price of the ship to be purely cost based, there would have been no reason for the act to have permitted the shipyard and the potential ship purchaser to *862 negotiate the ship's price. Rather, the act would have simply required the ship purchaser and the shipyard to agree on the design of the ship, and then the act would have directed the shipyard to determine, in consultation with MarAd, the actual cost of constructing the ship. It is apparent, therefore, that the act contemplates a shipyard sometimes having to justify a predetermined market-based ship price with cost estimates. Because the market-based price of a ship may sometimes be higher than the ship's cost-based price, the act presupposes that the cost estimates that a shipyard submits in its proposal support to justify a market-based price may not be the same as the estimates the shipyard would use to justify a cost-based ship price. Thus, the fact that a shipyard has not provided MarAd with the shipyard's subjective best estimates of how much constructing the ship will cost, does not mean that the shipyard has failed to satisfy the act's reporting requirements. As long as the cost estimates are objectively reasonable, the shipyard apparently has complied with the act.[9]

The statute does not give shipyards unbridled discretion as to what they may include in the proposal support they give to MarAd. First of all, shipyards may not submit proposal support that contains information which is factually incorrect, e.g., if a shipyard has ordered a ship's boiler for $100, the shipyard may not state in its proposal support that the ship's boiler cost $200. Moreover, though a shipyard may structure cost estimates to justify a previously determined price, the estimates must be objectively reasonable or the estimates will not satisfy the act, and MarAd must reject them. Thus, our holding today does not permit shipyards to submit arbitrary or false information as part of proposal support to MarAd.

In the case before us, we have found that all of the factual data that QSD provided to MarAd as part of GD's proposal support was accurate, and that all of the cost estimates that QSD provided to MarAd as part of the proposal support were fair and reasonable. Thus, the proposal support that QSD submitted to MarAd satisfies the requirements of the act.

Assuming, arguendo, that the act requires a shipyard to submit to MarAd the cost estimates that reflect the shipyard's subjective best estimates of the actual costs of constructing the ship, the Government's action against GD based on the proposal support GD provided to MarAd must still fail. As already found, the Government has not proven by a preponderance of the evidence that GD failed to submit to the Government what GD considered were GD's subjective best estimates. Specifically, we have found, inter alia, the Government's assertions that GD's best estimates were contained in the "LNG Master Planning Assignment," the data QSD submitted to the GD board of directors, and the data that GD submitted to Arthur Andersen, to be unconvincing. See supra.

We have further found that, whether analyzed objectively or subjectively, the proposal support GD provided to the Government in support of GD's CDS application was entirely proper and was not fraudulent. Accordingly, GD's proposal support cannot advance the Government's claim of liability against GD for mistake, unjust enrichment, and fraud, or render GD liable to the Government under the False Claims Act.

The Government asserts that GD had a duty to update its proposal support until July 26, 1977, the date the Maritime Subsidy Board approved GD's CDS's. We disagree. We believe that the Merchant Marine Act only required GD to keep the proposal support current, complete, and accurate until November 18, 1976, the date *863 that GD and MarAd's Division of Domestic Costs reached an agreement on LNG ship prices.

In the TINA context, a Government contractor is only obligated to keep its cost or pricing data complete, current, and accurate up to the date that the contractor and the Government contracting officer ("GCO") agree on the price of the contract. See 10 U.S.C. § 2306a(d) (1) (B) (under TINA "defective cost or pricing data are cost or pricing data which as of the date of agreement on the price of the contract ... were inaccurate, incomplete, or noncurrent.") (emphasis added); Conrac Corp., 74-1 BCA ¶ 10,605 at 50,292 (1974) (specifically recognizing that under TINA a Government contractor has "no duty to disclose data that first came into its possession after the date [the contractor certifies the correctness of the contract price] and prior to the date of contract execution.")[10]; Paceco, Inc., 73-2 BCA ¶ 10,119 at 47,560 (1973) (same); see also Aydin Monitor Systems, 83-1 BCA ¶ 16,500 at 81,997 (1983) (holding that under TINA "[t]he requirement for submission of cost or pricing data is met when all accurate cost or pricing data reasonably available to the contractor at the time of the agreement on price is submitted."). Thus, by analogy[11], GD was only obligated to update its proposal support until November 18, 1976, when the Division of Domestic Costs and GD agreed on the LNG ship prices.

The Government maintains that TINA cases are distinguishable from the instant case because the GCO with whom the contractor negotiates in the TINA context has the authority to bind the Government to agreements. By contrast, MarAd's Division of Domestic Costs does not have such authority, and can only recommend to the Maritime Subsidy Board that the Board approve a CDS application. We do not find this distinction to be material. In Paceco, Inc., 73-2 BCA ¶ 10,119 (1973), a TINA case, the Board of Contract Appeals held that after the GCO and a contractor have concluded price negotiations, and the contractor has certified to the GCO that the contractor's cost and pricing data are correct, the Government contractor has no obligation to provide the Government with subsequent vendor quotations. Id. at 47,559-60. This is so, even though at the time the GCO and the contractor conclude price negotiations, the Government has not yet formally awarded the contract to the contractor, and is therefore not yet formally and finally obligated to pursue in all circumstances the transaction with the contractor. Thus, analogously, even though MarAd was not bound by the price agreement until the Maritime Subsidy Board approved GD's CDS application, the record was for support purposes in all respects complete, and the mere lack of formal Board approval did not in any way affect QSD's duty to update its proposal support.[12]

Besides the analogy to TINA, the only other sources that the parties point to as aids in determining CDS reporting requirements are two sets of proposed, but unenacted regulations. These unenacted regulations are over twenty years old, and have never, in any reported case, been relied upon by the Government or anyone else for that matter. Because of the failure of the Departments of Commerce and Transportation[13]*864 to formally adopt these regulations, we do not find the updating requirements set out in them to have been mandatory on GD. In United States v. Helmsley, 941 F.2d 71 (2d Cir. 1991), cert. denied, ___ U.S. ___, 112 S. Ct. 1162, 117 L. Ed. 2d 409 (1992), the Government prosecuted Leona Helmsley for, inter alia, tax evasion. Citing a proposed, but never enacted treasury regulation which stated that the real property depreciation rate for tax deduction purposes was 7% per year for the first 10 years and 6% for the next 5 years, Helmsley argued that she had actually overpaid her taxes because she depreciated her real property only 6.67% per year. Id. at 84-85. The Court held that because the regulation was proposed but was never adopted, it was not mandatory, and that therefore, Helmsley's reliance on it was misplaced. See id. at 88; see also Mearkle v. Commissioner of Internal Revenue, 838 F.2d 880, 883 (6th Cir.1988) ("[P]roposed regulations are not entitled to the same deference as are final regulations. In the instance of a proposed regulation, the promulgating agency has not had the benefit of administrative hearings or of comments from interested persons concerning the advisability of modifying the proposed regulation or adopting it as final."); Fairfax Nursing Center, Inc. v. Califano, 590 F.2d 1297, 1300 (4th Cir.1979) ("Regulations which are proposed but not subsequently promulgated indicate little about [an] agency's statutory authority."). Similarly, in the instant case, the updating requirements set out in the proposed regulations were not mandatory on GD, and thus GD cannot be held liable if we determine that GD did not abide by them.[14]

Even were we to look to the proposed regulations as indicators of the act's updating requirements for CDS proposal support data, the regulations would still not be persuasive. The 1971 proposed regulations state that "[a]t the conclusion of the negotiations, the shipyard shall certify that the cost or pricing data which it has furnished is current, complete, and accurate." JX132 at P018495 (emphasis added). Thus, the 1971 proposed regulations support the notion that the shipyard's requirement to provide current, complete, and accurate data to MarAd ends at the time that the shipyard's price negotiations with the Division of Domestic Costs end. The 1972 proposed regulations provide, however, that the "applicant shall file ... any amendments necessary to keep all the information contained or furnished in connection with a pending application current and correct." Construction Differential Subsidy, 37 Fed. Reg. 6759, 6763 (1972) (emphasis added). Thus, the 1972 proposed regulations can arguably be read to suggest that even after the Division of Domestic Costs and the shipyard have agreed on price, the shipyard must continue to update the cost information until the contract is formally "signed and sealed." As is apparent, the proposed 1971 regulations and the proposed 1972 regulations contradict each other.[15] We believe that this conflict diminishes the proposed regulations' already meager interpretive usefulness, and that therefore we are *865 only left with the TINA analogy as an aid in interpreting the act's updating requirements for CDS proposal support data.[16]

We also believe that the practical requirements of having a settled record before the Board based upon a binding and closed negotiation in matters so routinely complex and time sensitive as that before us, precludes the reading of the statute proposed by the Government.

In sum, therefore, we hold that GD's obligation to keep the proposal support it submitted to MarAd current, complete, and accurate extended only until November 18, 1976, the date upon which the shipyard and the Division of Domestic Costs agreed on the domestic LNG ship prices.

II. The Government's Statutory and Common Law Claims with Respect to the Frigitemp Kickbacks are Preempted by the Anti-Kickback Act and therefore Must Be Dismissed

In the case before us, the Government has sued GD under the False Claims Act, 31 U.S.C.A. § 3729, et seq. (1983), and federal common law for damages stemming from the Frigitemp kickbacks, and the Government has also sued GD under federal common law for recision and restitution. The Government has not sued GD for damages under what used to be known as the "Anti-Kickback Act," 41 U.S.C.A. 51 et seq. (1965), because prior to the act's 1986 amendments, which apply prospectively only, see Anti-Kickback Enforcement Act of 1986, Pub.L. No. 99-634 § 3(a), 100 Stat. 3523, 3526 (Nov. 7, 1986), the Government could only recover kickback damages from a subcontractor or the recipient of a kickback, and not from a prime contractor. See H.R.Rep. No. 964, 99th Cong., 2d Sess. 10, 14-15 (Oct. 7, 1986), reprinted in 1986 U.S.Code Cong. & Admin.News 5960, 5967, 5971-72. The Government concedes that GD did not receive a kickback, and GD maintains that, even though the Government could not have successfully sued GD under the Anti-Kickback Act because GD is a prime contractor, the Anti-Kickback Act preempts all of the Government's statutory and federal common law remedies against GD with respect to the Frigitemp Kickbacks. GD contends, therefore, that all of the Government's Frigitemp kickback claims must be dismissed as to GD. We agree.

First of all, the Supreme Court has quoted favorably "the general principle that `a precisely drawn, detailed statute preempts more general remedies.'" Jett v. Dallas Independent School District, 491 U.S. 701, 734, 109 S. Ct. 2702, 2722, 105 L. Ed. 2d 598 (1989) (plurality opinion) (quoting Brown v. General Services Administration, 425 U.S. 820, 834, 96 S. Ct. 1961, 1968, 48 L. Ed. 2d 402 (1976)). The precisely drawn, detailed statute Congress passed as a remedy for the injuries suffered by the Government as a result of kickback schemes is the Anti-Kickback Act. Thus, in the case before us, where the Government is suing GD for the harm the Government suffered as a result of the Frigitemp kickbacks, the Anti-Kickback Act, under this principle, preempts the False Claims Act and federal common law remedies.

Moreover, the Supreme Court has held that when Congress states in a statute's legislative history that, outside that statute, Congress knows of no remedy to cure the ill which the statute was passed to cure, Congress' statement indicates that Congress intended the statute to create an exclusive remedy which preempts all other facially applicable remedies. See Jett v. Dallas Independent School District, 491 U.S. at 734-35, 109 S.Ct. at 2721-22; Brown v. General Services Administration, 425 U.S. 820, 825-29, 96 S. Ct. 1961, 1964-66, 48 L. Ed. 2d 402 (1976). In Brown, the plaintiff, an employee of the General Services Administration ("GSA"), filed a complaint with the GSA Equal Employment Opportunity Office claiming that he had been denied a promotion because he was an *866 African American. Id. at 822, 96 S. Ct. at 1963. After an investigation and a hearing, the GSA finally determined "that considerations of race had not entered into the promotional process." Id. at 823, 96 S. Ct. at 1963. GSA then "told Brown that if he chose, he might carry the administrative process further by lodging an appeal with the Board of Appeals and Review of the Civil Service Commission and that, alternatively, he could file suit within 30 days in federal district court." Id.

Forty-two days later, Brown filed suit in federal district court claiming that the GSA violated 42 U.S.C.A. § 1981 and 42 U.S.C.A. § 2000e-16 ("§ 717") when GSA failed to promote him. Id. at 823-24. Brown's § 717 claim was time barred because he did not file his action within 30 days of GSA's final action on this matter. See 42 U.S.C.A. § 2000e-16(c). However, Brown apparently maintained that the Court could not dismiss his § 1981 employment discrimination action. See Brown, 425 U.S. at 824-25, 96 S. Ct. at 1963-64. Citing legislative history of § 717 which showed that when Congress passed § 717 it believed that there existed no other remedy for discrimination in federal employment, the Supreme Court held that by enacting § 717 Congress intended "to create an exclusive preemptive administrative and judicial scheme for the redress of federal employment discrimination." Id. at 829, 96 S. Ct. at 1966. The Court noted that the question of whether Congress' perception of the law was correct when Congress passed § 717 was unimportant because "the relevant inquiry is not whether Congress correctly perceived the state of the law, but rather what its perception of the state of the law was." Brown, 425 U.S. at 828, 96 S. Ct. at 1966.

In Jett v. Dallas Independent School District, 491 U.S. 701, 109 S. Ct. 2702, 105 L. Ed. 2d 598 (1989) (plurality opinion), the Supreme Court revisited the question of how courts can determine whether Congress intended a statutory scheme to preempt other federal remedies. In Jett, the plaintiff, Norman Jett, was relieved from his position as football coach of South Oak Cliff High School. Jett sued the school's principal and the Dallas Independent School District ("DISD") under 42 U.S.C.A. §§ 1981, 1983 claiming, inter alia, that he had been relieved from his coaching position because he was white. Id. at 705-07, 109 S. Ct. at 2706-08. One of Jett's theories about DISD's § 1981 liability was that DISD was liable for the principal's actions under the doctrine of respondeat superior. See id. at 707, 109 S. Ct. at 2707. The Supreme Court disagreed.

After reviewing the legislative history of § 1981, which was silent on the issue of respondeat superior liability, the Court noted that it had held in the past that municipalities cannot be held vicariously liable under § 1983 for the actions of their officials. See id. at 711-29, 109 S. Ct. at 2709-19. The Court asserted that the legislative history of § 1983 indicates that the 42nd Congress believed that by passing § 1983 it was enacting "the only[] federal damages remedy for the violation of federal constitutional and statutory rights by state Governmental actors." Id. at 734, 109 S. Ct. at 2722. The Court also pointed out that the legislative history of § 1983's 1874 revisions "further indicates that Congress thought that the declaration of rights in § 1981 would be enforced against state actors through the remedial provisions of § 1983." Id. Relying on this legislative history, and citing Brown, the Court held "that the express `action at law' provided by § 1983 for the `deprivation of any rights, privileges, or immunities secured by the Constitution and laws,' provides the exclusive federal damages remedy for the violation of rights guaranteed by § 1981 when the claim is pressed against a state actor." Id. at 735, 109 S. Ct. at 2722.[17] The *867 Court therefore rejected Jett's assertion that DISD could be held liable under § 1981 based on respondeat superior liability.

Thus, under the reasoning of Jett and Brown, the appropriate inquiry to use in determining whether Congress intended a federal statute to preempt all other federal remedies, is whether Congress indicated in the statute's legislative history that the statute was the only remedy for the ill which Congress passed the statute to cure.

When Congress first passed the Anti-Kickback Act in 1946, it stated unequivocally that "[t]here is no existing statutory or other authority of law under which it may be said that the United States clearly has a right to recover the amounts of any such [kickback] fees or gratuities." H.R.Rep. No. 212, 79th Cong., 1st Sess. (1945), reprinted in 1946 U.S.Code Cong. & Admin.News 1081, 1083; see also H.R.Rep. No. 1880, 86th Cong., 2d Sess. (1960), reprinted in 1960 U.S.Code Cong. & Admin.News 3292, 3293 (letter of the Comptroller General of the United States to the Speaker of the House of Representatives stating that the reason Congress passed the Anti-Kickback Act in 1946 was that "there was no specific statutory remedy for recovery by the Government of the amount of the kickback, and there was serious doubt as to the adequacy of common law remedies."); S.Rep. No. 1585, 86th Cong., 2d Sess. 2 (1960) (stating that Congress passed the Anti-Kickback Act in 1946 because "there was no specific statutory remedy for recovery by the Government of the amount of the kickback and there was serious doubt as to the adequacy of common law remedies.") Moreover, when Congress amended the Anti-Kickback act in 1960, it justified passing the amendment by pointing to, inter alia, a letter the Comptroller General of the United States sent to the Attorney General of the United States. The letter discussed a particular kickback episode and stated: "We know of no Federal statute under which [those involved in the kickback activities] could be held amenable to the United States in a criminal or civil action for fraud unless the transaction in question may be said to come within the purview of the Anti-Kickback Act." H.R.Rep. No. 1880, 86th Cong., 2d Sess. (1960), reprinted in 1960 U.S.Code Cong. & Admin.News 3292, 3297. Thus, it appears from the legislative history of the Anti-Kickback Act that Congress understood that the act was the only remedy the Government had for injuries the Government suffered as a result of kickback schemes.

Under the pre-1986 Anti-Kickback act, the Government could recover as kickback damages only the value of the kickback involved from the subcontractor who gave the kickback or from the recipient of the kickback. See 41 U.S.C.A. § 51 (1965); S.Rep. No. 435, 99th Cong.2d Sess. 14 (1986). However,

kickbacks often end up costing the Government more than the amount of the kickback that is passed along through the contract. In addition to increased prices the Government may suffer increased costs from the delivery of substandard goods or by poor performance under the contract. Further, the Government incurs expenses in investigating and prosecuting kickback cases.

H.R.Rep. No. 964, 99th Cong., 2d Sess. 15 (Oct. 7, 1986), reprinted in 1986 U.S.Code Cong. & Admin.News 5960, 5972; accord id. at 5966-67. One could argue that because the pre-1986 Anti-Kickback Act only provides as damages the value of the kickbacks, that the Anti-Kickback Act preempts the False Claims Act and federal common law remedies only to the extent that they are used to recover damages up to the value of the kickback. Under this theory, the Government could use the False Claims Act or common law remedies to recover the kickback damages that exceed the value of the kickback. We do not agree with this reasoning.

*868 First of all, acceptance of this argument leads to the absurd conclusion that in 1946 Congress believed that the Government could not use the False Claims Act and federal common law to recover damages up to the value of the kickbacks themselves, but that the Government could use these remedies to recover any kickback damages that exceeded the value of the kickbacks. Moreover, the 1960 Anti-Kickback Act legislative history makes clear that Congress was of the opinion at that time that there was no remedy at all that the Government could use to recover any kickback damages. Therefore, we conclude that in passing the Anti-Kickback Act in 1946, and in amending it in 1960, Congress believed that the Anti-Kickback Act was the sole remedy under which the Government could recover kickback damages, and that Congress intended the Anti-Kickback Act to preempt any other facially applicable remedies that could theoretically be used to recover kickback damages.

In sum, using the reasoning of Jett and Brown, we hold that the Government may not use the False Claims Act and federal common law to sue GD for the injuries the Government suffered as a result of the Frigitemp Kickback Scheme. Thus, the Government's claims against GD with respect to the Frigitemp kickbacks must be dismissed.

We recognize that some non-second circuit cases decided after Brown have held that the Government may use the False Claims Act to recover damages caused by kickback schemes. See United States v. Killough, 848 F.2d 1523, 1531-32 (11th Cir. 1988); United States v. Rotner, D.P.M., 1989 WL 63209 (E.D.Pa.1989); see also United States v. Kensington Hosp., 760 F. Supp. 1120, 1128 (E.D.Pa.1991) ("the Government need not show injury resulting from kickbacks in order to recover under the False Claims Act," but just needs to show that the kickbacks occurred). However, none of these cases discuss how Brown or Jett affects the interplay of the False Claims Act and the Anti-Kickback Act. Thus, we do not believe that these Courts were presented with or decided this issue, and we do not read these cases as standing for the proposition that the Anti-Kickback Act does not preempt the False Claims Act in actions for kickback damages. Cf. Pennhurst State School & Hosp. v. Halderman, 465 U.S. 89, 119, 104 S. Ct. 900, 918, 79 L. Ed. 2d 67 (1984) ("when questions of jurisdiction have been passed on in prior decisions sub silentio, this Court [will not] consider[] itself bound when a subsequent case finally brings the jurisdictional issue before us." (citations and internal quotation marks omitted)).

The Government raises a number of arguments against a finding that the Anti-Kickback Act preempts the False Claims Act and federal common law as remedies for the harm the Government suffered as a result of the Frigitemp kickback scheme. We find none of the arguments persuasive.

The Government first contends that a preemption finding is incorrect because it leaves the Government remediless against GD for the Frigitemp kickback damages. However, it is not this Court that leaves the Government remediless against GD for the Frigitemp kickback damages, it was Congress. Prior to 1986, Congress failed to make prime contractors liable under the Anti-Kickback Act. Moreover, the fact that Congress in 1986 amended the Anti-Kickback Act to provide for vicarious liability against prime contractors who pass along kickback costs, suggests that in 1986, Congress did not believe that the Government had any effective civil remedies to recoup kickback damages from prime contractors. See generally H.R.Rep. No. 964, 99th Cong., 2d Sess. 15 (Oct. 7, 1986), reprinted in 1986 U.S.Code Cong. & Admin.News 5960, 5972 (discussing amendment to the Anti-Kickback Act that "allow[s] the Government to sue a prime contractor or higher tier sub-contractor for showing payment or acceptance by a lower tier subcontractor.") Had Congress believed that such civil remedies existed, Congress would not have found it necessary to amend the act to include expressly vicarious liability. Thus, the legislative history of the successive enactments of Anti-Kickback Act from 1946 to 1986 consistently demonstrates that during this *869 40 year period Congress was of the opinion that there was no means for the Government to recoup kickback damages from prime contractors.

The Government next asserts that Brown is distinguishable from the instant case because § 717, the preemptive statute in Brown, contains "an elaborate non-judicial enforcement mechanism," that demonstrates that Congress intended § 717 to be the exclusive remedy for discrimination in federal employment. The Government maintains that the fact that the Anti-Kickback Act does not contain such a mechanism evidences that Congress did not intend the Anti-Kickback Act to be the exclusive remedy for the injuries the Government suffers as a result of kickback schemes. See Government's Brief in Response to GD's Preemption Argument at 12. We disagree.

Though it is true that the Brown Court pointed to § 717's elaborate remedial scheme as evidence that Congress intended § 717 to be the exclusive remedy for discrimination in federal employment, the Court did not hold that such an elaborate remedial scheme was necessary for a statute to be considered exclusive and preemptive. The Court merely relied on the scheme as an additional indication of § 717's exclusivity. See Brown, 425 U.S. at 829, 96 S. Ct. at 1966 ("We need not, however, rest our decision on [Congress' perception of the law, as reflected in the legislative history,] alone. For the structure of [§ 717] itself fully confirms the conclusion that Congress intended [§ 717] to be exclusive and pre-emptive."). That the presence of an elaborate remedial scheme is not necessary for a statute to be considered exclusive and preemptive becomes even clearer when one analyzes the Court's holding in Jett, supra. There, though 42 U.S.C. § 1983 did not contain a complex remedial scheme, the Supreme Court, relying on legislative history, held that "the express `action at law' provided by § 1983 ... provides the exclusive federal damage remedy for the violation of rights guaranteed by § 1981 when the claim is pressed against a state actor." 491 U.S. at 735, 109 S. Ct. at 2722; see id. at 731-36, 109 S.Ct. at 2720-23; id. at 748, 109 S. Ct. at 2729 (Brennan, J., dissenting) (pointing out that in the plurality's preemption analysis the plurality was not concerned with the fact that § 1983 does not have a complex remedial scheme). Thus, the Anti-Kickback Act's lack of an elaborate remedial scheme does not undermine our holding that the Anti-Kickback Act preempts all of the Government's other remedies for any injuries caused by kickback schemes.

The Government also contends that Brown is distinguishable from the instant case because Brown involved problems of sovereign immunity, see Brown, 425 U.S. at 833, 96 S. Ct. at 1968, while the instant case involves no such problems. It is true that, in distinguishing the case before it from a case in which the Court held that Title VII did not preempt other remedies for employment discrimination, the Brown Court pointed out, without elaboration, that the case before it implicated sovereign immunity concerns. However, the Court also relied on Title VII's legislative history to distinguish the cases. Moreover, Jett did not involve sovereign immunity concerns, yet the Court held § 1983 preemptive in that case. In short, we do not find the lack of sovereign immunity issues in this case as in any way affecting our holding that the Anti-Kickback Act preempts all of the Government's other remedies for harms caused by kickback schemes.

Citing United States v. Acme Process Co., 385 U.S. 138, 87 S. Ct. 350, 17 L. Ed. 2d 249 (1966), the Government claims that "[t]he Supreme Court has unequivocally rejected the argument that the sanctions enumerated in the Anti-Kickback Act were meant to be exclusive." See Government's Brief in Response to General Dynamics' Preemption Argument at 16. We disagree. In Acme Process, the Acme Process Equipment Company ("Acme") "undertook through itself and its subcontractors to manufacture [for the Government] 2,751 75-mm recoilless rifles for about $337 per rifle." Id. at 138, 87 S. Ct. at 352. Subsequently, the Government discovered that some Acme officials awarded subcontracts *870 to subcontractors in exchange for kickbacks, and the Government cancelled the Acme contract. See id. at 138, 141, 87 S. Ct. at 352, 353. Acme then brought a breach of contract action against the Government in the United States Court of Claims. Id. at 138, 87 S. Ct. at 352. The Government argued that because Acme officials had accepted kickbacks, the Government had the right to cancel the contract. Id. Acme countered that the Government did not have the power to cancel the contract because the Government was limited to the remedies provided for in the Anti-Kickback Act, which did not include contract cancellation. The Court disagreed with Acme and stated that the public policy against kickbacks expressed in the Anti-Kickback Act "require[d] that the United States be able to rid itself of a prime contract tainted by kickbacks." Id. at 146, 87 S. Ct. at 356; see id. at 143, 87 S. Ct. at 354 ("The Anti-Kickback Act not only `prohibited' [kickbacks] [], but clearly expressed a policy decidedly hostile to them. They were recognized as devices hurtful to the Government's procurement practices."). Thus, the Court's narrow holding in Acme Process is that the Government cannot be required to continue to perform a contract once the Government determines that the contract is infected by kickbacks. Acme Process does not, as is argued by the Government, stand for the proposition that the Government can use federal common law and statutes other than the Anti-Kickback Act as remedies for the injuries suffered by the Government as a result of kickback schemes.[18]

The Government next argues that because the Anti-Kickback Act's legislative history reflects that law enforcement personnel from various Government agencies informed Congress that they were able to obtain criminal kickback convictions with statutes other than the Anti-Kickback Act, we must assume that Congress did not mean the Anti-Kickback Act to be the "[g]overnment's only avenue of redress for subcontractor kickbacks." Government's Brief in response to General Dynamic's Preemption Argument at 14-15. However, all that this legislative history proves is that when Congress passed the Anti-Kickback Act it was aware that the criminal portion of the Anti-Kickback Act would not be the exclusive means for prosecutors to obtain criminal convictions for kickback schemes. This legislative history does not in any way contradict the fact that when Congress passed the Anti-Kickback Act in 1946, and amended it in 1960, Congress believed that it was creating the only civil remedy under which the Government could recover kickback damages. As the case before us is a civil suit, the legislative history about the Act's criminal provisions is inapposite.

We observe that to support its contention that the Anti-Kickback Act does not preempt the Government's federal common law and False Claims Act claims in this case, the Government has cited cases which reflect the traditional reluctance of federal courts to find state law preempted by federal statutes. See Government's Brief in response to General Dynamic's Preemption *871 Argument at 6-8. "Such reluctance [is] particularly appropriate in light of the Supreme Court's repeated emphasis on the central role of Congress in protecting the sovereignty of the states." L. Tribe, American Constitutional Law § 6-25 (1988). However, this federalism concern is inapplicable in cases, such as the one before us, where a court is analyzing preemption of federal law by a federal statute. Thus, the traditional reluctance to find preemption in a federal/state context does not apply in these circumstances.

Analogizing its role in this civil litigation to that of a federal prosecutor in a criminal case, the Government has cited a number of criminal cases which stand for the proposition that

absent evidence of Congressional intent to repeal, when a new statute overlaps a portion of an older one, the two statutes should be permitted to coexist unless the two are mutually exclusive.

United States v. Jackson, 805 F.2d 457, 461 (2d Cir.1986), cert. denied, 480 U.S. 922, 107 S. Ct. 1384, 94 L. Ed. 2d 698 (1987); see United States v. Batchelder, 442 U.S. 114, 118-122, 99 S. Ct. 2198, 2201-03, 60 L. Ed. 2d 755 (1979); Edwards v. United States, 312 U.S. 473, 484, 61 S. Ct. 669, 675, 85 L. Ed. 957 (1941); United States v. Bilzerian, 926 F.2d 1285, 1299-1300 (2d Cir.), cert. denied, ___ U.S. ___, 112 S. Ct. 63, 116 L. Ed. 2d 39 (1991); United States v. Jackson, 805 F.2d 457, 461 (2d Cir.1986); United States v. Shareef, 634 F.2d 679, 680-81 (2d Cir.1980). Relying on these cases, the Government argues that we should not read the Anti-Kickback Act as "repealing" the False Claims Act in suits for kickback damages. The Government's argument is meritless.

As we noted above, when Congress passed the Anti-Kickback Act in 1946, and amended it in 1960, Congress was of the opinion that there existed no federal law under which the Government could recover damages the Government suffered from kickback schemes. Thus, the entire question of whether Congress intended the Anti-Kickback Act to repeal the False Claims Act is inapplicable in the instant case because when Congress passed the Anti-Kickback Act, Congress did not believe that the False Claims Act overlapped any portion of the Anti-Kickback Act, and did not believe that the False Claims Act was a viable remedy in the kickback context. Moreover, as opposed to the case before us, and in the cases cited by the Government and referred to in the preceding paragraph, either Congress indicated in the legislative history of the newer statutes that the older statutes were to coexist with the newer statutes, see United States v. Batchelder, 442 U.S. at 119-20, 99 S.Ct. at 2201-02; United States v. Jackson, 805 F.2d at 463-64, or Congress had not made explicit its understanding that the newer statutes were the sole remedies for the ills the newer statutes were passed to cure. See Edwards v. United States, 312 U.S. 473, 484, 61 S. Ct. 669, 675, 85 L. Ed. 957 (1941) ("We see no basis for the conclusion that Congress intended to repeal the earlier statute. The two can exist and be useful side by side." (emphasis added)); United States v. Bilzerian, 926 F.2d 1285, 1299-1300 (2d Cir.) (If Congress does not make the indicia of repeal explicit, "the general rule that criminal statutes may overlap controls."), cert. denied, ___ U.S. ___, 112 S. Ct. 63, 116 L. Ed. 2d 39 (1991); United States v. Shareef, 634 F.2d 679, 680-81 (2d Cir.1980) (same). Therefore, we find unpersuasive the Government's arguments about repeal.

Accordingly, we hold that the Anti-Kickback Act preempts the Government's statutory and common law claims with respect to the Frigitemp kickbacks. Therefore, the Government's Frigitemp claims, must be, and are dismissed.


Having considered the full record and found the aforesaid facts, and having reviewed all of the relevant law, we conclude that all of the Government's claims against General Dynamics must be dismissed. The complaint is accordingly dismissed, and the Clerk of the Court is directed to enter judgment in favor of General Dynamics.



[1] Since 1981, the Secretary of Transportation rather than the Secretary of Commerce has handled CDS applications.

[2] References to "JX" are to joint exhibits.

[3] References to GD are to General Dynamic's exhibits.

[4] References to Tr. are to the minutes of the trial transcript.

[5] References to "GX" are to government exhibits.

[6] The Government also contends that GD's manning analyses, which the Government asserts were based on the CAC's contained in JX175, evidence that GD did not believe that its learning curve was 94%. Specifically, the government points out that the CAC's in JX175, which the government asserts GD never gave to MarAd, are different from the CAC's GD gave to MarAd in JX196. We do not believe that this discrepancy, by itself, establishes fraud. Though Renn stated at trial that this document was used in GD's manning planning, other testimony indicates that the purpose of this document was to "put pressure on the operations department to man to meet budget. They might not be able to accomplish that, but that would be the purpose of the exercise." See JX720 at 304 (McAndrew Dep.). In short, we find that this document, like the operating plan, contains aggressive assumptions that were based on operational goals rather than learning curve theory, and therefore we do not find that this document establishes fraud.

[7] We note that the government has also asserted a different damage calculation in GX395 and GX396. However, the government has provided no reason why we should prefer its damage estimate in GX395 and GX396 over its estimate in GX128-F, and GD looks to GX128-F to help it to determine the Frigitemp damages. Therefore, we also look to GX128-F to help us to determine the Frigitemp damages.

[8] Section I of the conclusions of law should be read as referring to all of GD's proposal support, except for GD's estimate for Frigitemp's joiner and sphere insulation work. The government's claims with respect to the Frigitemp joiner and sphere insulation work are analyzed in section II.

[9] Citing 1971 proposed, but never enacted, regulations which state that a CDS applicant should submit to MarAd "[a] copy of all detailed estimate backup sheets upon which the proposed price is based," JX514 at 1505 (emphasis added), the government argues that GD violated the Merchant Marine Act by submitting data that GD structured to justify a predetermined price. For the reasons discussed infra, we do not find the 1971 proposed regulations to have been binding on GD. Moreover, for the reasons discussed above, we believe that the act contemplates shipyards justifying market-based prices with structured cost estimates.

[10] The date of certification is apparently the same date upon which the price negotiations end.

[11] In construing a statute, Courts may look to constructions of analogous statutes as interpretive aids. See Grayned v. City of Rockford, 408 U.S. 104, 110, 92 S. Ct. 2294, 2300, 33 L. Ed. 2d 222 (1972) (To determine the meaning of a state statue which has not yet been interpreted by the state courts, federal courts may look to, inter alia, interpretations the state courts have given to analogous statutes); United States v. Deutsch, 451 F.2d 98, 110 (2d Cir.1971) (As aids in interpreting the Investment Company Act, the court looked to, inter alia, interpretations of statutes analogous to the act); see also United States v. Rose, 549 F. Supp. 830, 832 (S.D.N.Y.1982) ("[A]ssistance in interpreting statutory language may be drawn from the interpretation of similar language in earlier analogous legislation.")

[12] As noted above, the reasons the Maritime Subsidy Board's approval was delayed until July 26, 1977 had nothing to do with price.

[13] As noted above, until 1981 the MarAd functioned as part of the Department of Commerce. Since 1981, MarAd has functioned as part of the Department of Transportation.

[14] United States v. Regan, 937 F.2d 823 (2d Cir.1991), modified on other grounds, 946 F.2d 188 (2d Cir.1991), cert. denied, ___ U.S. ___, 112 S. Ct. 2273, 119 L. Ed. 2d 200 (1992), does not require a different result. In Regan, the Court stated that a proposed, but never adopted regulation "shed light on the meaning of [the statute in question]." Id. at 825. We do not read this case as requiring courts to accept a proposed regulation's interpretation of a statute. Rather, we believe that the Regan decision indicates that a Court may, but not must look to a proposed regulation as an interpretive aid. That Regan does not require courts to accept a proposed regulation's interpretation of a statute is evident from the fact that in the Second Circuit's subsequent decision in Helmsley, supra, the Court declined to do so. The majority was clearly aware of the Regan case from Judge Oakes' dissent, see Helmsley, 941 F.2d at 106 (Oakes J., concurring in part and dissenting in part), but the majority apparently read Regan in the way that we propose.

[15] However, if an application is not considered "pending," within the meaning of the 1972 regulations, after negotiations between the applicant and the Division of Domestic Costs have concluded, then the 1972 regulations do not contradict the 1971 regulations.

[16] Any argument that the 1972 proposed regulations somehow repealed the 1971 proposed regulations is meritless. Since both sets of regulations were never enacted, there is no reason to believe that the 1972 proposed regulations are any more authoritative than the 1971 proposed regulations.

[17] The Court also pointed out that a number of circuit courts had refused to infer from the Fourteenth Amendment a Bivens-type action under which municipalities could be held vicariously liable for the acts of their officers. Jett, 491 U.S. at 735, 109 S. Ct. at 2722. The Court recognized that the circuit courts had refused to create such a remedy because, inter alia, Congress had not seen fit to impose vicarious liability on municipalities when Congress passed § 1983, and there was no legal or logical basis for the imposition of such liability. Id. Stating that "the logic of these decisions applies with equal force to [Jett's] invitation, ... [the Court declined] to create a damages remedy broader than § 1983 from the declaration of rights [] found in § 1981." Id.

[18] We recognize that in Acme Process the Court stated that "[t]here is absolutely no indication in the legislative history of the Anti-Kickback Act that Congress, in providing a civil remedy for a more tangible evil, intended to preclude other civil sanctions necessary to effectuate the purpose of the Act." Id. at 145. We believe that, when read in context with the rest of the opinion, the "civil sanctions" referred to by the Supreme Court are defensive sanctions to breach of contract actions, such as the "sanction of nonenforcement" discussed in the paragraph immediately preceding the statement in question. Moreover, even if the Acme Process Court intended the words "other civil sanctions" to include actions for damages, recision, and restitution, we would still hold that the government may not sue GD for the Frigitemp kickbacks under the False Claims Act and federal common law. In the first place, the Court's statement about civil sanctions besides contract cancellation is nonbinding dicta as the case before the Court was limited to contract cancellation. Second, the Court's 1966 assertion that there is no evidence in the Anti-Kickback Act's legislative history that shows that Congress intended the Anti-Kickback Act to be preemptive, is puzzling. As noted above, when one examines the legislative history of the Anti-Kickback Act using the mode of legislative history analysis propounded in the Court's 1976 decision in Brown, and its 1989 decision in Jett, it becomes apparent that Congress did indeed intend that the Anti-Kickback Act be preemptive.