ATK Thiokol, Inc. v. United States

United States Court of Appeals for the Federal Circuit
2009-5036

ATK THIOKOL, INC. (now known as ATK Launch Systems Inc.),

Plaintiff-Appellee,

v.

UNITED STATES,

Defendant-Appellant.

Thomas A. Lemmer, McKenna Long & Aldridge LLP, of Denver, Colorado, argued
for plaintiff-appellee. With him on the brief was Phillip R. Seckman.

Robert E. Chandler, Trial Attorney, Commercial Litigation Branch, Civil Division,
United States Department of Justice, of Washington, DC, argued for defendant-appellant.
With him on the brief were Tony West, Assistant Attorney General, Jeanne E. Davidson,
Director, and Donald E. Kinner, Assistant Director. Of counsel on the brief was Douglas R.
Jacobson, Defense Contract Management Agency, Contract Disputes Center, of Ft.
Snelling, Minnesota. Of counsel was James W. Porier, Trial Attorney, of Washington, DC.

Stephen D. Knight, Smith Pachter McWhorter P.L.C., of Vienna, Virginia, for amicus
curiae Committee on Government Business of Financial Executives International.

David A. Churchill, Jenner & Block LLP, of Washington, DC, for amicus curiae
National Defense Industrial Association. Of counsel on the brief were Kevin C. Dwyer,
Jessie K. Liu, and Michael A. Hoffman.

Appealed from: United States Court of Federal Claims

Judge Susan G. Braden
 United States Court of Appeals for the Federal Circuit

2009-5036

ATK THIOKOL, INC. (now known as ATK Launch Systems Inc.),

Plaintiff-Appellee,

v.

UNITED STATES,

Defendant-Appellant.

Appeal from the United States Court of Federal Claims
in 99-CV-440, Judge Susan G. Braden.

__________________________

DECIDED: March 19, 2010
___________________________

Before BRYSON, ARCHER, and MOORE, Circuit Judges.

BRYSON, Circuit Judge.

I

ATK Thiokol, Inc., manufactures rocket motors for government and commercial

buyers. One of its products is the Castor family of solid rocket motors, which are “strap-

on” motors designed to attach to a launch vehicle and provide additional thrust. In the

1990s, ATK began developing the Castor IVA-XL rocket. In 1995, ATK announced that

it was closing the Huntsville, Alabama, plant where it had previously built the Castor

IVA-XL motor. At that time, ATK analyzed each of the products made by the Huntsville

plant as part of a corporate restructuring effort. ATK determined that there was a
market for the Castor IVA-XL motor and that the motor would be more competitive if it

were upgraded. Specifically, ATK decided to make technical changes to the motor’s

design and test fire the motor. The parties refer to those steps as the “Development

Effort.” ATK moved production of the Castor IVA-XL motor to its Utah facility in 1995.

From 1995 through 1999, ATK marketed the upgraded motor to various potential

customers, including McDonnell Douglas, Lockheed Martin, and the U.S. Air Force.

In February 1996, the Japanese company Mitsubishi Heavy Industries expressed

interest in purchasing modified Castor IVA-XL motors for Japanese government launch

vehicles. While Mitsubishi was willing to pay for adapting and attaching the motors to

the launch vehicles, it refused to pay for the more general, nonrecurring work

associated with upgrading the Castor IVA-XL motor, in particular the Development

Effort. On June 8, 1997, Mitsubishi executed a Statement of Work with ATK. The

Statement of Work required ATK to deliver the motor that ATK was updating “to support

the general requirements of the strap-on market.” ATK began the Development Effort in

July 1997. In October 1998 ATK and Mitsubishi executed a final contract for the

purchase. The contract provided for a lump-sum payment for each upgraded motor and

a price for modifying each motor to fit to the Japanese launch vehicles. There was no

provision for payment of the Development Effort costs.

ATK accounted for the Development Effort costs as indirect independent

research and development (“IR&D”) costs and disclosed them as such in a proposed

advance agreement submitted in 1997 to a U.S. Defense Department Divisional

Administrative Contracting Officer (“DACO”). The effect of including particular charges

as indirect IR&D costs is that the charges are allocated to all of the contractor’s

2009-5036 2
contracts for the year, including government contracts. ATK’s consistent and disclosed

accounting practice was to treat research and development costs as indirect costs

unless (1) the particular contract in question specifically required that ATK incur the

cost; (2) the contract paid for the cost; or (3) the cost had no reasonably foreseeable

benefit to more than one cost objective. From 1990 through 1997, the Defense

Department had periodically reviewed that accounting practice and found it to be in

accordance with the accounting regulations applicable to government contractors.

In March 1999, however, the DACO issued a notice of intent to disallow the cost

of the Development Effort for the upgraded Castor IVA-XL motor by removing that cost

from the category of indirect IR&D. The DACO noted that the definition of IR&D in the

Federal Acquisition Regulation (“FAR”) excludes efforts “required in the performance of

a contract.” The DACO reasoned that, because the upgrade cost was necessary to the

performance of the Mitsubishi contract, the Development Effort did not qualify as IR&D

and that it had to be charged, if at all, directly to the Mitsubishi contract.

ATK filed an action in the Court of Federal Claims, arguing that the DACO’s

refusal to treat the Development Effort costs as indirect IR&D was improper. The court

found that ATK and Mitsubishi did not intend to include the Development Effort costs

among the costs paid for under the contract, that the commercial market for the

upgraded Castor motor appeared viable, that the allocation of the Development Effort

costs to indirect IR&D was in accordance with ATK’s disclosed accounting practices,

and that the government had not contended that the Development Effort costs were

unreasonable. ATK Thiokol, Inc. v. United States, 68 Fed. Cl. 612, 640-41 (2005).

Based on those findings, and after reviewing the pertinent FAR provisions, the

2009-5036 3
corresponding sections of the Cost Accounting Standards (“CAS”), and the regulatory

history of those provisions, the court held that the Development Effort costs were

chargeable as indirect IR&D. Id. at 641. The government appeals from that decision.

II

Whether particular research and development costs qualify as indirect IR&D for

purposes of government contract accounting is determined by several interrelated

regulations. First, section 402 of the Cost Accounting Standards, 48 C.F.R. § 9904.402

(“CAS 402”), defines direct and indirect costs. A “direct cost” is “any cost which is

identified specifically with a particular final cost objective”; an “indirect cost” is “any cost

not directly identified with a single final cost objective, but identified with two or more

final cost objectives, or at least one intermediate cost objective.” CAS 402-30(a)(3), (4).

CAS 402 gives the contractor considerable freedom in the classification of particular

costs, so long as the contractor maintains consistency in making that determination.

See CAS 402-20; see also Boeing Co. v. United States, 862 F.2d 290, 292-93 (Fed. Cir.

1988).

Second, two parallel regulations determine whether certain costs qualify as

IR&D. A provision of the Federal Acquisition Regulation, 48 C.F.R. § 31.205-18 (“FAR

31.205-18”), determines whether particular costs are allowable IR&D charges. A

provision of the Cost Accounting Standards, 48 C.F.R. § 9904.420 (“CAS 420”),

determines whether those costs are allocable to the particular contract in question.

Both the FAR and the CAS define IR&D as excluding costs that are “required in the

performance of a contract.” FAR 31.205-18(a); CAS 420-30(a)(6).

2009-5036 4
 In light of the language and interpretation of CAS 402, it was appropriate for ATK

to treat the Development Effort costs at issue in this case as indirect costs. First, those

costs were not specifically required by the Mitsubishi contract. Second, as the trial court

found, ATK had a disclosed and established cost accounting practice of charging as

indirect costs those costs that were not paid for or required by a particular contract and

that had a reasonably foreseeable benefit to more than one contract.

While we conclude that it was proper for ATK to treat its Development Effort

costs as indirect, that does not end the inquiry, because the distinction between IR&D

and other research and development costs does not invariably track the distinction

between direct and indirect costs. For example, depending on a contractor’s disclosed

or established cost accounting practices, the contractor may treat some research and

development costs as indirect costs because they benefit an entire product line, even if

they are expressly required by a particular contract and thus would not qualify as IR&D.

See 1 Karen L. Manos, Government Contract Costs & Pricing § 25:6, at 396 (2009).

As the trial court’s analysis makes clear, however, the costs at issue in this case

qualify as IR&D costs. The dispute over that issue focuses principally on the meaning

of the phrase “required in the performance of a contract” in the definition of IR&D. The

scope of that phrase has been a subject of controversy in the government contracts

community since 1971, when it first appeared as part of the definition of IR&D in the

Armed Services Procurement Regulation (“ASPR”), a predecessor to the FAR. 1 The

government interprets the phrase to mean that IR&D costs do not include the costs of

1
In 1978, the ASPR was redesignated as the Defense Acquisition Regulation
(“DAR”). The DAR definition of IR&D costs was essentially the same. The FAR
replaced the DAR in 1984 without any change to the definition of IR&D.

2009-5036 5
efforts that are either explicitly or implicitly required in order to complete a contract.

Under that interpretation, the Development Effort costs would be deemed to be

“required in the performance of” the Mitsubishi contract because that contract could not

have been performed if ATK had not conducted the background research and

development work to upgrade the Castor IVA-XL motor. ATK, by contrast, interprets the

phrase to be limited to costs that are explicitly required by the contract. Under that

interpretation, the Development Effort costs would not be deemed “required in the

performance of” the Mitsubishi contract because there was no explicit requirement in

the contract that obligated ATK to undertake the Development Effort or provided for

payment of the Development Effort costs.

The government argues that the phrase that excludes costs from the category of

IR&D must be construed broadly because it does not simply exclude costs “required by

a contract,” but uses broader language, excluding costs “required in the performance of

a contract.” Thus, the government contends that the words “in the performance of a

contract” indicate that the focus is on whether the research and development work was

necessary to perform the contract, not whether the work was expressly required by the

contract. The government relies heavily on the district court decision in United States v.

Newport News Shipbuilding, Inc., 276 F. Supp. 2d 539 (E.D. Va. 2003), which adopted

that interpretation. ATK, on the other hand, contends that the term “required” must refer

to a requirement of the contract. According to ATK, if the drafters of the regulation had

intended to include research and development costs that were needed but not expressly

included as a contract requirement, they would have used a word such as “necessary”

rather than the word “required.”

2009-5036 6
 We do not find either textual argument particularly persuasive. Standing alone,

the language of the regulation is ambiguous. See United States ex rel. Mayman v.

Martin Marietta Corp., 894 F. Supp. 218, 222 (D. Md. 1995) (noting the “considerable

debate” over whether “a contractor can bill to IR&D any work not explicitly called for in

the contract” or whether the contractor may not bill to IR&D anything “implicitly

necessary to carry it out”); John W. Chierichella, IR&D vs. Contract Effort, in 90-2 Gov’t

Contract Costs, Pricing & Accounting Report 8 (Feb. 1990) (noting “sustained intra-

Governmental debate and confusion” over whether research and development effort not

specified or directly funded by a contract may be disallowed as IR&D because it is

deemed “implicitly” necessary).

Both parties invoke the regulatory history for support. Like the text, however, the

regulatory history is inconclusive. 2 Prior to the 1971 amendment that added the

disputed language, the applicable regulation defined IR&D as “research and

development which is not sponsored by a contract, grant, or other arrangement.” The

“ASPR committee,” which was responsible for adopting changes to the ASPR,

suggested changing the definition to cover “technical effort which is not sponsored by,

or in support of a contract, grant, or other arrangement.” Representatives of the

defense and space industries objected that the words “in support of” were too broad and

would be a potential “source for future misinterpretation.” The ASPR committee

acknowledged that the point was “a valid objection” and altered the language by

inserting “required in performance of” in place of the phrase “in support of.” The

2
The trial court’s opinion contains a thorough and well-documented review of
the regulatory history. 68 Fed. Cl. at 635-39. We have merely summarized the
pertinent points here.

2009-5036 7
committee explained that the purpose of the change was “to convey the concept that

any work which must be accomplished in order to fulfill contractual requirements is a

contract cost. Other similar type effort may be and should be expected to be

performed” as IR&D. The industry representatives then suggested further clarifying the

definition by adding the words “specifically required by contract provisions” before the

words “in performance of a contract.” The ASPR committee, however, rejected that

suggestion. The Comptroller General subsequently suggested that the exclusion from

the definition of IR&D was too narrow and that the exclusion should be amended to

include “technical effort implicitly required to fulfill a purchaser’s requirement under

terms of a contract.” A Department of Defense representative, however, took a position

contrary to that of the Comptroller General, urging that the exclusion from IR&D should

be understood to turn on “whether the effort was specified as a deliverable requirement

of an existing contract.” The ASPR committee declined to make any further change in

the pertinent regulatory language.

This sequence of events is not particularly instructive as to the meaning of the

regulatory definition. Although the ASPR committee declined the industry’s suggestion

to limit the exclusion from IR&D to costs “specifically required” by the contract, it also

declined the Comptroller General’s suggestion to broaden the exclusion to include costs

“implicitly required” by the contract. When the Board responsible for the Cost

Accounting Standards (“the CAS Board”) promulgated CAS 420 in 1979, it incorporated

in its definition of IR&D the same language that was used in the FAR—“required in the

performance of a contract.” The CAS Board did not offer any explanation as to the

2009-5036 8
intended meaning of that phrase, and there has been no change in the pertinent

regulatory language since then. The regulatory history is thus inconclusive.

While we find the regulatory language and history to be of little help in discerning

the meaning of the phrase “required in the performance of a contract,” we agree with

the trial court and ATK that the meaning of that phrase in the definition of IR&D must be

the same as the meaning of the identical phrase in the definition of bid and proposal

(“B&P”) costs. B&P costs are defined to mean costs incurred in preparing, submitting,

and supporting bids and proposals, but not to include the costs of effort “required in the

performance of a contract.” FAR 31.205-18(a); CAS 420-30(a)(2). B&P costs are

addressed in the same regulations that govern IR&D costs and are treated similarly to

IR&D costs in all pertinent respects. See generally FAR 31.205-18; CAS 420-30. B&P

costs “benefit all business of a contractor rather than a specific existing contract [and

thus] treating all such costs as indirect overhead is logical.” Boeing, 862 F.2d at 293.

A provision of CAS 402, referred to as Interpretation No. 1, supplies important

guidance as to when proposal costs constitute B&P and when they are chargeable

against a single contract. It provides, in pertinent part:

[C]osts incurred in preparing, submitting, and supporting proposals
pursuant to a specific requirement of an existing contract are considered
to have been incurred in different circumstances from the circumstances
under which costs are incurred in preparing proposals which do not result
from such specific requirements. The circumstances are different
because the costs of preparing proposals specifically required by the
provisions of an existing contract relate only to that contract while other
proposal costs relate to all work of the contractor.

CAS 402-61(c), 48 C.F.R. § 9904.402-61(c). Interpretation No. 1 distinguishes proposal

costs that are “specifically required by” an existing contract from those that “do not

result from such specific requirements.” The former costs “relate only to [a particular]

2009-5036 9
contract,” while the latter costs “relate to all work of the contractor” and thus qualify as

B&P. The effect of Interpretation No. 1 is to equate the B&P definitional exclusion of

proposal costs that are “required in the performance of a contract” with the category of

costs that are “specifically required by the provisions of a contract.”

In Boeing, this court held, based in part on Interpretation No. 1, that proposal

costs that are not specifically required by a contract are “properly allocated as indirect

B&P costs.” 862 F.2d at 293. As part of its analysis, the court noted that proposal costs

that are “specifically required by an existing contract are incurred in circumstances

different from proposal costs relative to all work of the contractor.” Id. The court ruled

that it would be legal error to require like accounting for B&P costs that are “relate[d] to”

or “caused or generated by” a contract, and those proposal costs that are “specifically

required” by a contract. Id. at 292-93.

The same analysis applies to the closely analogous category of IR&D costs.

Although Interpretation No. 1 does not by its terms address IR&D, the government’s

suggestion that the approach employed in Interpretation No. 1 should be limited to B&P

costs, and that IR&D costs should be treated differently, would result in a construction in

which identical regulatory language—“required in the performance of a contract”—would

be interpreted differently for IR&D than for B&P. There is no support anywhere in the

text or history of the regulations for treating that identical regulatory formulation

differently. We therefore construe the reference to costs “required in the performance of

a contract” to mean, in both contexts, costs that are specifically required by the contract.

The government argues that we should be skeptical of accepting such a result

because it is contrary to sound procurement policy. In particular, the government

2009-5036 10
argues that allowing a government contractor to charge to an indirect IR&D account

those research costs that are necessary to complete a commercial contract but not paid

for in that contract will invite a contractor to “game the system” by shifting commercial

contract costs to the government.

We are not persuaded by the government’s policy argument. First, the purpose

of IR&D is to benefit both government contractors and their customer agencies by

encouraging the contractors to engage in research that is likely to benefit multiple

contracts, both governmental and commercial. Spreading IR&D costs across multiple

contracts encourages general research that enables the contractor to innovate, to

maintain a high level of technological sophistication, and ultimately to improve the

products it offers the government. As the Department of Defense has explained,

providing financial support for IR&D serves several Departmental goals, including

creating “an environment that encourages DoD contractors to expand knowledge in

mathematics and science, improve technology in areas of interest to the Department of

Defense, and enrich and broaden the spectrum of technology available to the

Department of Defense.” Dep’t of Def. Directive No. 3204.1, at 3 (May 10, 1999).

Second, the result of requiring IR&D costs to be borne by a contract for which the

research and development work in question is deemed necessary could have the

perverse effect of charging all of the research and development costs for a proposed

product line against the first contract for the products in that line, whether the contract is

governmental or commercial. That approach would either disproportionately burden the

contract that happened to be first in line or ensure that the first contract would be a

losing one. For research that, by hypothesis, benefits multiple potential contracts, both

2009-5036 11
commercial and governmental, allocating general research and development costs in

that manner is not sensible as a policy matter.

Accordingly, the government’s policy arguments do not persuade us that the

phrase “required in the performance of a contract” in the definition of IR&D costs should

not be accorded the same meaning as the identical phrase in the definition of B&P

costs. Because the research and development costs at issue in this case were related

to the Mitsubishi contract but were not specifically required by that contract, we uphold

the trial court’s decision that those costs were indirect IR&D costs within the meaning of

the pertinent regulatory provisions.

AFFIRMED.

2009-5036 12