Smithkline Corporation v. Elmer Staats, Comptroller General of the United States, and the United States of America

OPINION OF THE COURT SEITZ, Chief Judge. SmithKline Corporation appeals from a final judgment of the district court granting the Government’s motion for summary judgment and holding that the Comptroller General (the Comptroller) has the right of access to records of costs “of materials, labor, overhead, distribution, research and development and marketing, and such other costs as are directly pertinent to the drugs purchased” under several contracts between the United States and SmithKline. SmithKline Corp. v. Staats, 483 F.Supp. 712, 722 (E.D.Pa.), conditional petition for cert. before judgment denied, 449 U.S. 1038, 101 S.Ct. 618, 66 L.Ed.2d 501 (1980). This court has jurisdiction under 28 U.S.C. § 1291 (1976). I. During 1973 and 1974, SmithKline entered into five negotiated contracts 1 to supply drugs to the Veterans Administration and the Defense Supply Agency. As required by statute, each contract contained an access-to-records provision granting the Comptroller access to “any directly pertinent” records “involving transactions related to this contract.” See 10 U.S.C. § 2313(b) (1976) (Defense Supply Agency); 41 U.S.C. § 254(c) (1976) (Veterans Administration). *203 In 1971, Senator Gaylord Nelson, chairman of a subcommittee investigating the pharmaceutical industry, learned of the access-to-records provisions contained in certain government contracts and urged that the General Accounting Office (GAO) invoke the provisions: “I think that we ought to take a look at some of those costs.” Competitive Problems in the Drug Industry: Hearings on Present Status of Competition in the Pharmaceutical Industry Before the Subcomm. on Monopoly of the Senate Select Comm, on Small Business, 91st' Cong., 2d Sess. 8020 (1971). One of Senator Nelson’s staff assistants, Ben Gordon, explicitly stated that specific cost and price data should be secured by individual company and product, and that “such data should be made public.” Gordon demanded on behalf of Senator Nelson that the data be sought “through the courts if necessary,” and “without any strings attached so that the information could be used as needed.” Similarly, a staff assistant to Senator Edward Kennedy, chairman of a Senate subcommittee on health, insisted that the “only way” the-objectives of Senator Kennedy’s subcommittee could be achieved was to “publicize specific price and cost data for individual products.” GAO finally concluded that it had “no viable alternative than to press the companies for access to their cost data.” Initially, however, GAO sought the voluntary participation of drug manufacturers. Its appeal was successful; SmithKline and five other manufacturers voluntarily participated in the first phase of a proposed two-phase study. In April 1974, GAO issued a proposal for “Phase II,” designed to “gather and develop the data necessary” to make a “presentation concerning salient economic and operational aspects of the industry.” The study was to cover research and development, manufacturing, marketing, distribution, and corporate overhead. GAO requested data relevant to all these aspects of the companies’ business for the period between 1964 and 1973. SmithKline expressed reservations about giving GAO access to such cost and pricing data without adequate guarantees of confidentiality. The Comptroller was prepared to give such guarantees in the interests of completing the study. Between April and August of 1974, GAO representatives met with Senate staff to seek their consent to the assurances of confidentiality. GAO emphasized that it would obtain “far more data” from voluntary disclosure, obtained on the basis of guaranteed confidentiality, “than we could demand under our statutory authority.” The staff members were unpersuaded and refused to sanction any assurances of confidentiality. On August 26, 1974, abandoning the concept of a voluntary study, the Comptroller issued formal demand letters, seeking access to the records of the companies that had participated in the Phase I study. This step was taken with the approval of Senator Nelson. GAO was under no legal obligation to accede to the Senators’ demands, and the demands had not been formally made by their respective subcommittees. GAO’s sole reason for abandoning its voluntary program and resorting to formal demands was the Senators’ refusal to. approve any assurances of confidentiality. GAO had no reason to believe that any impropriety had occurred in connection with the contracts, that any excess or unfair profits had been made, or that any violation of the law or other irregularity had occurred. All of the contracts selected by GAO for demand letters were fixed-price procurements, with prices established not by actual negotiation, but by reference either to catalogue prices for ordinary commercial items sold in significant quantities to the general public or to other evidence of a competitive price. SmithKline filed this action seeking a declaratory judgment that the Comptroller lacked the authority to require the production of the data demanded. 2 The Govern *204 ment counterclaimed for declaratory and injunctive relief to compel SmithKline to comply with the Comptroller’s demand. The district court granted the Government’s motion for summary judgment, holding that GAO shall have access to all of SmithKline’s cost records that the Government contended were directly pertinent to the price of the drugs sold under the contracts. II. SmithKline makes two main arguments on appeal: (1) that GAO sought information from SmithKline for an improper purpose; and (2) that the scope of the demand exceeds that allowed by the access-to-records provisions. The improper purpose argument raises three distinct issues: (a) whether GAO exceeded its statutory authority to audit the performance of particular Government contracts by demanding access to confidential financial records for the purpose of conducting a general study of procurement techniques or of conducting an economic survey of the profitability of the pharmaceutical industry; (b) whether GAO exceeded its statutory authority by demanding access to confidential financial records, at the instance of two Senators and their staff aides, in order to furnish the Senators with individual company and product data through the conduct of an economic survey of the profitability of the pharmaceutical industry; and (c) whether the district court erred in granting summary judgment because the court’s interpretation of the record created a material factual issue regarding the actual purpose of GAO’s de-' mand. SmithKline’s arguments are premised on the principle that an agency that is given power to investigate for one purpose cannot use that power to pursue distinctly different, unauthorized goals. See, e.g., United States v. LaSalle National Bank, 437 U.S. 298, 307, 316-17 n.18, 98 S.Ct. 2357, 2367-68 n.18, 57 L.Ed.2d 221 (1978) (IRS subpoena power); United States v. Garden States National Bank, 607 F.2d 61, 68 (3d Cir. 1979) (same); United States v. Humble Oil & Refining Co., 488 F.2d 953, 954 (5th Cir. 1974) (holding that the IRS exceeded its authority in issuing a summons in furtherance of a project for which the primary purpose was research rather than the investigation and adjustment of tax liabilities), vacated and remanded, 421 U.S. 943, 95 S.Ct. 1670, 44 L.Ed.2d 97, reaff’d, 518 F.2d 747 (5th Cir. 1975); Burlington Northern, Inc. v. ICC, 462 F.2d 280, 286 (D.C.Cir.) (ICC request for an injunction), cert. denied, 409 U.S. 891, 93 S.Ct. 120, 34 L.Ed.2d 148 (1972). Assuming, without deciding, that the contractual powers of GAO, a unique independent agency within the legislative branch, are subject to the same restraints as those on the subpoena powers of administrative agencies, 3 we proceed to examine whether the undisputed facts support SmithKline’s assertion that GAO sought access for an improper purpose. A. SmithKline argues that GAO exceeded its limited authority to audit the performance of particular contracts under 10 U.S.C. § 2313(b) and 41 U.S.C. § 254(c) because it demanded access to confidential records for the stated purpose of conducting a general study of procurement techniques. The critical premise of such a position is that Congress enacted the access-to-records provisions to prevent profiteering and fraud, and not to sanction a general scrutiny of pro *205 curement techniques or an economic survey of a particular industry. According to SmithKline, Congress sought to deter specific abuses and not to permit general investigations of the records of companies which are not in any way suspected of fraud or impropriety in their dealings with the Government. The provisions were certainly intended to authorize access to records in cases of suspected fraud and profiteering. See, e.g., 97 Cong.Rec. 13,200 (1951) (remarks of Rep. Riehlman); 96 Cong.Rec. 17,123 (1951) (remarks of Rep. Celler). Nevertheless, given that such was the primary concern, it does not necessarily follow that no other purposes for the provisions exist. There is a fine line between profiteering and wastefulness, and the vast majority of the legislative references to the provisions do not indicate an intent to restrict their use to situations involving impropriety. The problems of noncompetitive pricing detailed in the congressional debates include waste in Government procurement. Representative Hardy, the sponsor of the access-to-records provisions, summarized the congressional purpose as follows: “The major purposes of this bill are twofold: One, to give the Comptroller General the proper tools to do the job the Congress has instructed him to do; and, two, to provide a deterrent to improprieties and wastefulness in the negotiation of contracts.” Id. at 13,198. Representative Hardy further explained that examination of contractors’ records by the Comptroller would enable “the General Accounting Office to look behind the [price] which had been established [to assess whether it was] excessive” and would act as a “safeguard against waste and extravagance” in Government procurement. Id. Cf. Schwegmann Brothers v. Calvert Distillers Corp., 341 U.S. 384, 394-95, 71 S.Ct. 745, 750-51, 95 L.Ed. 1035 (1951) (sponsor’s statement of purpose is persuasive indication of congressional intent). Representative Hoffman remarked in the course of the debates that “GAO is the most effective agency we have against fraud and extravagance and waste and all those things.” 97 Cong.Rec. 13,373 (1951). See id. at 13,199 (remarks of Rep. Hardy) (Congress “clearly understood ... that there are a lot of other situations besides those involving .fraud which might be uncovered”). SmithKline points to passages in the legislative history indicating that Congress intended only to detect fraud and profiteering with the access provisions. See id. at 13,373 (remarks of Rep. Hardy) (reassuring the House that exercise of the provisions would not be routine but that GAO could invoke them if it has “reason to suspect fraud or bad faith or illegality” in connection with the contract); id. at 12,611 (remarks of Rep. Hardy) (stating the “sole purpose of the bill is to enable the Comptroller General to make effective audits along the lines and under the jurisdiction conferred upon the General Accounting Office by law”). We are not convinced, however, that these scattered statements are indicative of the thrust of the legislative history. The most reasonable construction of the access-to-records provisions is that Congress has authorized GAO to uncover waste by invoking them. We also find support for our interpretation of the statute and its legislative history in the case law. See Merck & Co. v. Staats, 665 F.2d 1236, 1237 (D.C.Cir.1981) (Mikva, J., concurring in part and dissenting in part); United States v. Abbott Laboratories, 597 F.2d 672, 674 (7th Cir. 1979); Eli Lilly & Co. v. Staats, 574 F.2d at 915-16. But cf. Note, Government Contracts: Contractor’s Obligation to Allow Examination of Records Under 10 U.S.C. § 2313(b), 72 Dickinson L.Rev. 687, 691 (1968) (stating that access provisions were intended “solely to provide a method to discover fraudulent activities,” but recognizing that “there is nothing in the Act specifically negating” the possibility that Congress sought to improve general procurement techniques). Although the statute and its legislative history do not specifically indicate that Congress intended to authorize GAO studies of procurement techniques, they authorize GAO to seek information that will help the Government reduce wastefulness and that *206 is “directly pertinent” to the contract. Acquiring data for studies of procurement techniques seems a reasonable means to the end that Congress intended. In light of our own review of the statutes and their legislative history, we are unpersuaded that GAO’s interpretation of the access-to-records provisions as allowing such access is unreasonable.. See Hewlett-Packard v. United States, 385 F.2d 1013, 1016 (9th Cir. 1967) (noting that, if production costs are “out of line with the contract price, the contract may have been an inappropriate means of meeting this particular procurement need.. .. [and this approach] could lead to the use of other methods of meeting future procurement needs”), cert. denied, 390 U.S. 988, 88 S.Ct. 1184, 19 L.Ed.2d 1292 (1968). Great weight is traditionally accorded to the construction of a statute by the agency charged with its administration “unless there are compelling indications that it is wrong.” Red Lion Broadcasting Co. v. FCC, 395 U.S. 367, 381, 89 S.Ct. 1794, 1801, 23 L.Ed.2d 371 (1969). SmithKline points to evidence that GAO has expressly and impliedly indicated, in letters to congressional committees at the time the statute was enacted, in internal memoranda, and in recommendations that Congress grant it additional authority, that it lacks the power to examine records relating to nongovernment business activities or to conduct “should cost” studies. SmithKline also contends that GAO has never previously conducted economic “industry” studies. At the most, SmithKline’s evidence suggests that GAO has changed its interpretation. More likely, GAO’s interpretations represent GAO’s search to discover which tasks are reasonably related to its responsibility for overseeing Government purchases. In any case, we believe that GAO has correctly concluded that it has the statutory authority to make the studies that it has proposed. Nor do we think it determinative that Congress has only once granted GAO a right of access to records of government contractors for the purpose of conducting an economic study, or that in 1973 and 1975 Congress declined to act on several proposals drafted by GAO that would have granted GAO access to contractors’ records in order to conduct profit studies of Government contractors. See, e.g., American Trucking Associations v. Atchison, Topeka & Santa Fe Railway, 387 U.S. 397, 418, 87 S.Ct. 1608, 1619, 18 L.Ed.2d 847 (1967) (agency’s request for legislation is not regarded as administrative interpretation of enabling statute); United States v. UMW, 330 U.S. 258, 281-82, 67 S.Ct. 677, 690, 91 L.Ed. 884 (1947) (subsequent legislative history declaring intent of an earlier statute not controlling). Relying on the fact that Congress did not pass a bill to interpret a bill that Congress passed many years earlier is inherently suspect. Such inaction certainly does not indicate that Congress itself has concluded that the provisions themselves do not grant authority for the studies. The issue here is whether the access-to-records provisions now before us can reasonably be read to permit GAO’s present demand. We conclude that they can. Assuming that GAO actually seeks SmithKline’s records for the purpose of conducting an economic study of the pharmaceutical industry, we believe that it is a small step to find that the Comptroller’s audit responsibilities include the study not only of the firms that contract with the Government but also the industries of which the firms are a part. GAO must have some understanding of an industry before it can study an individual procurement effectively. The Comptroller’s Phase II demand in this case is equivalent to a request for access to records relevant to a cost-price comparison of each contract for the purpose of determining whether there has been “wastefulness” in pharmaceutical procurement. That GAO will facilitate its comparison by compiling an industrywide economic study or procurement techniques study does not vitiate that purpose. B. SmithKline also argues that the record clearly demonstrates that GAO seeks access to SmithKline’s confidential business *207 records to undertake an economic study of the pharmaceutical industry in response to intense and continuing pressure from two Senators and their staff aides who wish to obtain, for public disclosure, cost and pricing data relating to specific products and manufacturers. 4 Assuming, as we must, for purposes of reviewing the district court’s summary judgment, that SmithKline could prove that GAO is seeking to obtain information for an industrywide economic study that would provide Senators Nelson and Kennedy with the information they sought, we cannot hold that GAO’s purpose is improper. The statutory requirement that the access-to-records provision be included in the contracts establishes the public interest in uncovering wastefulness and extravagance. That United States senators encourage and influence GAO to use its powers to the fullest extent allowed by lav.', and that GAO heeds those senators, is irrelevant. See Parnell, Congressional Interference in Agency Enforcement: The IRS Experience, 89 Yale L.J. 1360, 1368 (1980), quoting Gravel v. United States, 408 U.S. 606, 625, 92 S.Ct. 2614, 2627, 33 L.Ed.2d 583 (1972) (noting that “members of Congress ‘are constantly in touch with the Executive Branch of Government — they may cajole, and exhort with respect to the administration of a federal statute’ ”). We believe that a demand letter issued by GAO pursuant to its statutory authority, under the continuing and intense pressure of united States senators, may not, without more, be resisted on the ground that GAO sought the information for an improper purpose. In SEC v. Wheeling-Pittsburgh Steel Corp., 648 F.2d 118 (3d Cir. 1981) (en banc), we indicated that “we will not countenance judicial interference with agency decisions to conduct investigations, decisions that a>e committed entirely to agency discretion. ...” Id. at 127 (citing, e.g., City of Chicago v. United States, 396 U.S. 162, 165, 90 S.Ct. 309, 311, 24 L.Ed.2d 340 (1969)). 5 The reason for this policy of noninterference was recently stated by the United States Court of Appeals for the Fifth Circuit: “Judicial supervision of agency decisions to investigate might hopelessly entangle the courts in areas that would prove to be unmanageable and would certainly throw great amounts of sand into the gears of the administrative process.” Dresser Industries, Inc. v. United States, 596 F.2d 1231, 1235 n.1 (5th Cir. 1979), cert. denied, 444 U.S. 1044, 100 S.Ct. 731, 62 L.Ed.2d 730 (1980). We believe those same concerns support our result in this case. Moreover, although the courts that have considered whether GAO may investigate fraud and whether it was motivated by a proper purpose in investigating the pharmaceutical industry have taken an approach different from our own, see Lilly, 574 F.2d at 910-12 (finding speculation about senatorial prodding unnecessary if GAO otherwise has a *208 legitímate purpose); Merck, at 1239-1240 (Mikva, J., concurring in part and dissenting in part) (same); SmithKline, 483 F.Supp. at 720-21, no court has found that GAO is precluded from demanding records from pharmaceutical firms because of an improper purpose. C. SmithKline argues that at a minimum this court must decide that material issues of fact regarding the actual purpose of the Comptroller exist, thus rendering summary judgment for the Government inappropriate. We have concluded, however, that, assuming GAO sought to conduct a broad economic study of the pharmaceutical industry in order to provide cost and pricing data to individual Senators, GAO did not invoke the access-to-records provisions for an improper purpose. Thus, SmithKline’s contention is without merit, because the issue as to the Comptroller’s actual purpose is not material to our legal conclusion. III. SmithKline- also alleges that the district court erred in its determination of which records may be considered “directly pertinent” to “transactions relating to” the contracts. SmithKline makes two arguments as to the scope of access: (1) cost records are pertinent only to cost-based contracts; and (2) even if some cost data is pertinent to fixed-price contracts, the district court failed to set any reasonable limits on the scope of GAO’s access. A. The five contracts at issue are fixed-price contracts. The prices were established by reference either to catalogue prices for ordinary commercial items sold in significant quantities to the general public or to other evidence of competitive price. SmithKline argues that cost data cannot be “directly pertinent” to “transactions related to this contract” where the price is fixed not in reference to a standard of fairness in negotiating or to costs, but to catalogue, market, or competitive prices. In support of this interpretation of sections 2313(b) and 254(c), SmithKline points to limitations in the Truth-in-Negotiations Act, 10 U.S.C. § 2306(f) (1976), and in the Renegotiation Act of 1951, 50 U.S.C.App. § 1216(e) (1976), on the submission of cost data for the negotiation or renegotiation of contracts based on established catalogue prices or adequate price competition. SmithKline argues that these statutes and the regulations promulgated thereunder provide important evidence that similar restrictions are incorporated in sections 2313(b) and 254(c) through the language limiting access to material “directly pertinent” to “transactions relating to” the contract. SmithKline contends that, in the absence of convincing evidence that different interpretations were intended, interrelated statutory provisions should be given a harmonious construction. See Kokoszka v. Belford, 417 U.S. 642, 650, 94 S.Ct. 2431, 2436, 41 L.Ed.2d 374 (1974). SmithKline maintains that in this area of Government procurement Congress has consistently recognized that the presence of competition is a sufficient substitute for other safeguards against contractor abuse. Both the statutory and the contractual provisions clearly apply by their terms to negotiated contracts without qualification. If Congress had desired to exclude fixed-price contracts from the provisions’ operation, it could have explicitly done so, as it did in the Truth-in-Negotiations Act and in the Renegotiation Act. See Lilly, 574 F.2d at 916; Hewlett-Packard, 385 F.2d at 1016. Moreover, SmithKline’s contention that the access-to-records provisions do not extend to costs conflicts with the language of the provisions. Not only has SmithKline contracted to disclose some “directly pertinent” records in its possession, but it has also contracted to put a provision in its subcontracts securing access for GAO to the “directly pertinent” records of its subcontractors. If the position now urged by SmithKline that the statute and contract do not allow access to any cost records is correct, the provision for access to the records of SmithKline. and of its subcontractors would be a nullity. The purpose of the *209 access-to-records provision is to allow GAO to assess the reasonableness of the contractors’ costs and pricing by insuring full access by the Comptroller to all directly pertinent cost and pricing records, including subcontractors’ records that .otherwise might be unreachable because they would not be in the custody and control of the contractor. Thus, SmithKline’s reading ignores a large portion of the provision itself and violates the well-established rule that a contract and a statute should be read so as to give meaning to all their provisions. SmithKline argues that, because the access-to-reeords provisions are dictated by statute, they cannot be individually tailored to meet the specific requirements of each contract. The Government, SmithKline contends, can take no comfort from the fact that the access clauses seem to envisage the production of subcontractor records that appear to be relevant to the issue of costs and pricing. This clause is required by the statute whether or not SmithKline planned to use subcontractors in the performance of the particular contract. SmithKline argues that similarly the clause is required whether the subcontractor’s performance will be “directly pertinent” to the particular contract or not. Congress has by statute required a provision in every contract that in some instances has no application because the contractor does not employ a subcontractor. The possibility that a standard contractual provision will have no application in one contract does not, however, render it meaningless in another contract where it could have application. Thus, we think that cost records can be directly pertinent to fixed-price contracts and that an access-to-records provision in such a contract gives GAO the authority to request directly pertinent data. See Lilly, 574 F.2d at 916. B. Determining that the statutory provision does not exempt fixed-price contracts does not necessarily mean that all costs that affect the price of the drug covered in the contract are directly pertinent to the contract. Representative Hardy used the following examples to illustrate the situations which his bill was designed to reach: It is common practice today to provide in a contract for construction work a fixed overhead rate, either as a proportion of direct labor costs or as a lump sum amount. ... If in the course of checking the accounts of the accountable officer, the GAO found a situation indicating that the amount of overhead charged under the Government' contract was a great deal higher than the amount paid under a private contract for the same work, it would then obviously be desirable for the General Accounting Office to look behind the rate which had been established.... During the last war, there were cases where contractors held both fixed-price and cost-plus-a-fixed-fee contracts covering similar items or services, both contracts being performed simultaneously and in the same plant. In several cases it was found that the contractor was charging to the cost-plus-a-fixed-fee contract expenses which should have been charged to the fixed-price contract. In that way he was being reimbursed for expenses which should have been paid out of his own pocket. Such a practice may, or may not, have been intentional. But in any event, there would be little or no chance of detecting such a practice, unless the General Accounting Office has access to the contractors’ books and records. 97 Cong.Rec. 13,198 (1951). For GAO to detect a differential allocation of overhead costs between governmental and nongovernmental contracts, or between cost-plus and fixed-price contracts, it would have to examine a very broad set of records. See Case Comment, supra, at 1157-58 n.67. The implication of Representative Hardy’s examples is that the legislation was intended to authorize inquiry not only into all costs associated with each negotiated government procurement contract, but also into all similar contracts between a particular contractor and his private buyers, in order to determine whether the government *210 was bearing an excessive share of the company’s fixed costs. It is clear, however, that Congress did not intend the “directly pertinent” language to allow GAO to examine records of costs recovered exclusively from nongovernmental contracts. See Merck, at 1249-1250 (Mikva, J., concurring in part and dissenting in part); Memorandum on the Adequacy of the Legal Authority of the General Accounting Office to Conduct a Comprehensive Study of Profitability in Defense Contracting, reprinted at 115 Cong.Rec. 25,800 (1969) [hereinafter GAO Memorandum] (GAO has taken the position “that the words ‘directly pertinent’ were intended only to limit GAO’s right of access to records pertaining to Government work as distinguished from non-government work”). The bill that contained sections 2313(b) and 254(c) originally limited the Comptroller to “pertinent” records of the contractor. Representative Hoffman emphasized in debates over the access-to-records provisions that he objected strongly to “snooping expeditions” and that the access-to-records provisions would allow GAO “unlimited authority to go everywhere and snoop into everybody’s business.” 97 Cong.Rec. 13,373 (1951) (also stating opposition to a statute that “would let GAO go into everybody’s business and look it over”). To prevent an expansive interpretation of the provisions, Congress restricted the access proposal even further, by specifying that only records “directly” pertinent to the contract may be demanded. Representative Hoffman, who sponsored the limiting amendment but opposed the original bill, see 97 Cong.Rec. 13,377 (1951), stated specifically that the purpose of the modifying language was to “limit the ‘snooping’ that may be carried on under this bill.” Id. Representative Hoffman emphasized that, unless the proposed bill were limited, even a small government contract would open all of a contractor’s records to inspection by GAO. He pointed out that, without amendment, the provisions would permit a contract for ten percent of a contractor’s production to be inspected by GAO “not only with reference to that one item of 10 percent .. . but into all of his other business where that same item was used.” Id. He feared this result because the contractor would inevitably fail to keep “a separate set of books” for his government business, so that: “every section of his books will have to come under the complete scrutiny of the GAO. He is going to be harassed no end to try to break down production costs and show the details. ...” Id. at 13,376 (remarks of Rep. Harvey). Although these statements were offered in support of a proposed amendment relating to subcontractors, we believe that the same apprehension stimulated the introduction and adoption of the limiting “directly pertinent” language. The Government argues that Congress’ addition of the word “directly” did not sharply narrow the scope of inquiry to be allowed GAO. See Case Comment, supra, at 1157-58. But see Note, Eli Lilly & Co. v. Staats: An Undue Expansion of the GAO’s Investigatory Power Under the Access-to-Records Statutes, 74 Nw.U.L.Rev. 122, 135 (1979) [hereinafter Nw. Note], It is true that Representative Hardy did not oppose the amendment. See 97 Cong.Rec. 13,377 (1951). It is also true that Representative Hoffman did not indicate how his change would affect GAO’s ability to secure specific documents. See id. On the other hand, a •sponsor’s statement of purpose generally is regarded as a persuasive indicator of congressional intent. See Federal Energy Administration v. Algonquin SNG, Inc., 426 U.S. 548, 564, 96 S.Ct. 2295, 2304, 49 L.Ed.2d 49 (1976). In any case, we do not believe that the insertion meant nothing just because there was little debate on the change. The legislative history manifests a strong congressional intent to prevent unlimited “snooping.” “[W]hen Congress enacts a statute designed to limit government intrusion in the private affairs of its citizens, the statutory provisions must be followed scrupulously.” United States v. Bacheler, 611 F.2d 443, 447 (3d Cir. 1979). We thus conclude that the contract does not provide for unlimited access to SmithKline’s cost records. *211 When cost information is not allocated to specific contracts and their products, the Government’s argument that generalized cost data are “directly pertinent” to any given contract is weakened. See Nw. Note, supra, at 132. Like most manufacturing businesses, SmithKline divides its costs into a small category of direct manufacturing costs that can be readily assigned to a particular product and into other categories of cost, including research and development, distribution, marketing, advertising, and administrative costs, which are not assigned to any one product. See Reekie, Price and Quality Competition in the United States Drug Industry, 26 J.Indus.Econ. 223, 235 (1978); Nw. Note, supra, at 132-33. Cf. National Association of Cost Accountants, Research Series No. 29, Accounting for Research and Development Costs, 36 N.A.C.A. Bull. 1375, 1427-30 (1955) (survey finding that some businesses assign research and development costs to particular products, but some do not because the arbitrary method of allocation would not be helpful). SmithKline maintains that this common practice reflects the lack of generally accepted accounting principles that permit accurate and meaningful allocation of these indirect costs. Obviously, however, these unassigned costs affect the price charged for the company’s products, because the company must recover both the costs directly assigned and the unassigned pooled costs. Manufacturing and distribution costs of individual pharmaceutical products may constitute as little as nine percent of the products’ sale price. See Lilly, 574 F.2d at 913; Rucker, Public Policy Considerations in the Pricing of Prescription Drugs in the United States, 4 Int’l J. Health Services 171, 173 (1974). By characterizing unallocated costs as “directly pertinent” to government procurement contracts, GAO would be able to demand access to all records of pooled costs whenever the price charged the Government in the contracts at issue reflects costs attributable to the pool. GAO would thus have access to most, if not all, of Smith-Kline’s pricing data. 6 The Government, on the other hand, rejects both the argument that there are no generally accepted accounting principles that permit meaningful allocation of these indirect costs and SmithKline’s suggestion that GAO’s interpretation of the access provision would be without limits. The Government argues that SmithKline itself allocates costs to products and performs profitability studies for its own purposes; even if allocation of indirect costs to products cannot be made with scientific accuracy, it is helpful and necessary. The Government contends that there is almost no limit to the allocation of costs to particular products other than that set by expediency. The Government maintains, however, that it will not need certain information or information breakdowns and thus that its request for information is not without bounds. In short, the Government’s contention is that any cost affecting the prices charged the Government under a given contract may be “directly pertinent” to the contract. See Lilly, 574 F.2d at 914-15 (finding that an item is directly pertinent to a contract if it is a “significant input” in the price of the product purchased in the contract). We believe that the Government’s argument undermines its own position, for it would seem that the Government’s interpretation of “directly pertinent” admits of no doctrinal limitation on its access powers. We are unwilling to read the statutory provision as granting to GAO such broad discretion in determining what is pertinent and what is “snooping.” The Government has not suggested a definition of “directly pertinent” that preserves the distinction between snooping and properly limited access. Indeed, its argument gives the phrase no effect in this case. If a “directly pertinent” *212 cost record means any cost having a significant input into price, it would seem that GAO could also justify an extensive investigation of subcontractors who may actually be doing no Government business at all. Cf. Morgan, The General Accounting Office: One Hope for Congress to Regain Parity of Power with the President, 51 N.C. L.Rev. 1279, 1364-65 (1973) (“The same argument that would justify a GAO audit in Hewlett-Packard would justify an audit of contractors doing no government business at all.”). We do not think that an interpretation of the “directly pertinent” language that allows practically unlimited investigation of companies that do no direct business with the Government can be sound in light of the congressional desire to prevent snooping. The Government contends that GAO’s administrative interpretation of “directly pertinent” should be followed in the absence of “compelling indications that it is wrong,” especially where those interpretations were reported to Congress. We believe, however, that Congress’ inaction in the face of the report should not be read as acquiescence. See Girouard v. United States, 328 U.S. 61, 69, 66 S.Ct. 826, 829, 90 L.Ed. 1084 (1946). GAO’s broad interpretation of the provisions is inconsistent with the congressional intent and is entitled to no deference. Cf. Nw. Note, supra, at 135 n.77 (“The GAO itself has been uncertain about the exact limits of its authority under the access-to-records statutes.”). We must therefore search for a more limited interpretation if “directly pertinent” is to be given an interpretation consistent with the congressional history. In some industries, costs such as research and development have such a small impact that they arguably are not “directly pertinent” to the contract and its price. See Lilly, 574 F.2d at 913-14. Whether a cost record is “directly pertinent” could vary from industry to industry. See Merck, at 1249-1250 (Mikva, J., concurring in part and dissenting in part). In some industries, direct costs are the predominant component of a product’s price, and there would be no need to grant access to indirect cost information of those firms. But in the pharmaceutical industry, research and other costs not immediately attributable to one product form a large part of the costs of a pharmaceutical product. Each case could be decided on its facts by addressing one question: whether the records sought will show production costs and pricing of the contracts sufficient to determine why the contract price was fixed as it was and whether “the contract may have been an inappropriate means of meeting this particular procurement need of the Government.” Hewlett-Packard, 385 F.2d at 1016. The judicial decision as to what records are necessary would thus be essentially an accounting problem which is not unlike those problems which are common in day-to-day discovery involving corporations. See generally United States v. IBM Corp., 66 F.R.D. 186 (S.D.N.Y.1974); Fed.R.Civ.P. 33(c). Although it is tempting to take the opportunity to make the decision as equitably as possible in each case, we are not convinced that Congress intended for the interpretation of the access-to-records provision to depend on the facts of each ease. The same provision appears in most Government contracts. See Nash, Risk Allocation in Government Contracts, 34 Geo.Wash.L.Rev. 693, 694 n.3 (1966) (finding that only one quarter of all Government procurement contracts are advertised procurements). We believe that Congress intended the words to acquire some common meaning so that the Government and the contracting firm could know within reason what they were contracting to perform. In contracts involving many millions of dollars, many firms, and many products, certainty is more important than deciding each ease on its facts. In Bristol Laboratories Division of Bristol-Myers Co. v. Staats, 428 F.Supp. 1388 (S.D.N.Y.1977), aff’d per curiam, 620 F.2d 17 (2d Cir. 1980), aff’d by evenly divided court, 451 U.S. 400, 101 S.Ct. 2037, 68 L.Ed.2d 343 (1981), the district court equated the directly pertinent costs with “manufacturing costs (including raw and packag *213 ing materials, labor and fringe benefits, quality control and supervision); manufacturing overhead (including plant administration, production planning, warehousing, utilities, and security); royalty expenses; and delivery costs.” 428 F.Supp. at 1389. See id. at 1391. Specifically excluded, however, were data with respect to “research and development, marketing and promotion, distribution and administration (except to the extent that these areas may be included in the factors above).” Id. at 1389-90. The Government has not put forward an alternative definition of “directly pertinent” records that we should adopt if we do not agree with its interpretation of the statute. We therefore adopt the standard formulated in Bristol, which for the most part relies on the distinction between direct and indirect costs. We believe that it comports with Congress’ intent to preclude GAO’s scrutiny of nongovernmental contracts. Moreover, our distinction between information that is “directly pertinent” and that which is not follows the distinction drawn by every circuit to consider the issue with the exception of the Seventh Circuit. 7 See Merck, 529 F.Supp. 1, 2-3 (D.D.C.1977), aff’d per curiam, 665 F.2d 1236 (D.C.Cir.1981). The Government argues that GAO’s interpretation that the access-to-records provision provides for the access it seeks here has been well-known and judicially established since the Hewlett-Packard decision in 1967. We do not believe, however, that the Ninth Circuit sanctioned access to all indirect overhead cost data. The Ninth Circuit held that GAO should have access to “production costs.” 385 F.2d at 1016. Although the court never defined these costs, the district court’s order, affirmed by the court of appeals, referred specifically to “costs of direct material-, direct labor and overhead costs.” Id. Our decision includes direct costs and some overhead within its definition of “directly pertinent.” How much overhead the Ninth Circuit intended to include is uncertain. See Merck, at 1248 n.27 (Mikva, J., concurring in part and dissenting in part) (“Later decisions have disagreed over the Hewlett-Packard holding on this point.”). Compare Lilly, 574 F.2d at 913, with Bristol, 428 F.Supp. at 1391. Thus, Hewlett-Packard does not necessarily suggest a result contrary to that in our case. We have concluded that GAO may demand access to detect wastefulness and may make certain studies to help it in its task. We are aware, however, that the limiting interpretation that we have given the access-to-records provision may make it more difficult in some cases for GAO to obtain the information it may need to uncover wastefulness. 8 Economic studies based on such restricted data may be of limited utility, and if the access provisions were meant to equalize the relationship between the Government and private contractors, our reading of the clause fails to advance fully that purpose in this case. See Merck, at 1246-1247 (Mikva, J., concurring in part and dissenting in part). The difficulty, however, arises because (1) the stat *214 ute attempts to further the two conflicting congressional goals of limiting “snooping” and permitting broad access; and (2) the pharmaceutical industry presents special characteristics. The conflict in goals appears on the face of the statute itself. Cf. Nw. Note, supra, at 139 (“there can be no accommodation of the competing interests; either the GAO gets all of the information or it gets none of it”). For example, the fact that the demand is limited by the access-to-records provisions, to data within three years of final payment, see 10 U.S.C. § 2313(b), no doubt hampers the detection of wastefulness and the usefulness of any studies that are prepared. Congress also restricted GAO’s investigatory powers by not making the access-to-records provisions applicable to advertised as well as negotiated contracts. Information on producers doing no business with the Government would undoubtedly also be useful in the overall improvement of Government procurements, since the data could be compared with the data collected from companies contracting with the Government; but GAO cannot obtain that information under the provisions. See Morgan, supra, at 1364-65. We do not believe that we are free to ignore these limitations in trying to assure that GAO has meaningful access to the contractors’ records. GAO has itself recognized its dilemma: “While GAO’s legal authority would permit it to perform some of the work necessary in making a profit study .. . ., to do a meaningful study of profitability . . ., legislation should be enacted broadening its right of access to records . . .. ” GAO Memorandum, supra, reprinted at 115 Cong.Rec. 25,-801 (1969). “In establishing the ‘directly pertinent’ standard as a limitation on the GAO’s access rights, Congress apparently was willing to forego obtaining all the information that might be related to a government contract if the only alternative was to require a contractor to make available virtually all of its records.” Nw. Note, supra, at 134-35 (footnote omitted). If “directly pertinent” information is not sufficient to make the Comptroller’s investigation of certain industries complete, or if certain industries should be treated differently because of unique circumstances, Congress, and not this court, is better able to decide which information GAO needs to fulfill Congress’ expectations. SmithKline must allow GAO personnel to inspect its records with respect to manufacturing costs, manufacturing overhead, royalty expenses, and delivery costs. Smith-Kline need not produce data with respect to research and development, marketing and promotion, distribution and administration, except to the extent that these costs may be included in the manufacturing costs. IV. The order of the district court will be vacated and the case remanded for entry of an order consistent with this opinion. Footnotes [1] In negotiated procurements, the Government discusses the terms of the procurement with one or more contractors, with the contract going to the candidate whose proposal is most advantageous to the Government. Negotiated contracts are to be distinguished from advertised procurements, which award contracts on the basis of competitive bids submitted by contractors. See Case Comment, The Comptroller General’s Authority to Examine the Private Business Records of Government Contractors: Eli Lilly & Co. v. Staats, 92 Harv.L.Rev. 1148, 1148 n.1 (1979). [2] This case is one of several recent challenges to the Comptroller’s authority to examine the records of pharmaceutical companies pursuant to the access-to-records provisions. See United States v. Abbott Lab., 597 F.2d 672 (7th Cir. 1979); Eli Lilly & Co. v. Staats, 574 F.2d 904 *204 (7th Cir.), cert. denied, 439 U.S. 959, 99 S.Ct. 362, 58 L.Ed.2d 351 (1978); Merck & Co. v. Staats, 529 F.Supp. 1 (D.D.C.1977), aff’d per curiam, 665 F.2d 1236 (D.C.Cir.1981); Bristol Lab. Div. of Bristol-Myers Co. v. Staats, 428 F.Supp. 1388 (S.D.N.Y.1977), aff’d per curiam, 620 F.2d 17 (2d Cir. 1980), aff’d by evenly divided court, 451 U.S. 400, 101 S.Ct. 2037, 68 L.Ed.2d 343 (1981). [3] See Eli Lilly & Co. v. Staats, 574 F.2d 904, 909 (7th Cir.), cert. denied, 439 U.S. 959, 99 S.Ct. 362, 58 L.Ed.2d 351 (1976); Cibinic & Lasken, The Comptroller General and Government Contracts, 38 Geo.Wash.L.Rev. 349, 350 (1970) (GAO operates “on the hazy borderline between legislative and executive powers”). Cf. Case Comment, supra, at 1154-56 (discussing whether GAO is an agency). [4] We do not understand SmithKline’s argument to be that, assuming GAO lawfully has acquired information — under the pressure from the Senators — for the purpose of preparing an industrywide economic study, it cannot turn it over to individual Senators. Rather, we believe that SmithKline’s argument is simply that “the central purpose underlying the Comptroller General’s demands was to obtain exploitable cost information, desired by two Senators and their aides, through the conduct of a broadranging study of the economics of the pharmaceutical industry.” We therefore need not address what limits, if any, exist on GAO’s ability to disclose information obtained by invocation of the access-to-records provisions. The parties have agreed to provisions that substantially protect the confidentiality of the records. See SmithKline, 483 F.Supp. at 723-24 (confidential information may only be disclosed if: (1) identification of SmithKline as the source of the information is impossible; (2) if a committee or chamber of Congress requests the information; or (3) if a subpoena is issued for the information). [5] In Wheeling-Pittsburgh, this court recognized that the enforcement of an SEC subpoena may be challenged where the agency is abusing the court’s process. See 648 F.2d at 125. Although this case was called to counsel’s attention at oral argument, SmithKline does not contend that this case is controlling in the present circumstances. Indeed, SmithKline does not allege that GAO is acting in bad faith in seeking access to the information, Cf. id. at 125 n.9 (defining agency bad faith as “a conscious decision by an agency to pursue a groundless allegation without hope of proving that allegation”). [6] If GAO were allowed access to this information, it would presumably have to make a similarly arbitrary allocation. Thus, denying or granting access to such records will have little effect on GAO’s ability to evaluate the contractor’s pricing practices on a particular contract. See Nw. Note, supra, at 134. The information would, however, be useful in determining the extent of differential allocation between contracts. [7] We note that in Abbott Laboratories another panel of the Seventh Circuit addressed precisely the same issues as those addressed in Lilly. The district court had granted summary judgment to the Government on the basis of Lilly, and the circuit court affirmed. Two of the three judges, however, specifically expressed reluctance and concurred in the affirmance solely because they considered themselves bound by the earlier decision. See 597 F.2d at 674 (Pell, J., concurring); id. at 675 (Wood, J., concurring). [8] In industries where direct costs are the prime component of price, where firms are producers of only one product, or where the firms sell almost everything they make to the Govemment, our decision will have little limiting effect on GAO’s ability to make various studies. With other firms, manufacturing cost information alone might be useful for two purposes: (1) it might help to uncover fraud; and (2) where the products are part of a competitive market, which most fixed-price cost contracts, by definition, are, comparison of direct cost information could help the Government determine which company is assigning more overhead to its product. Nevertheless, we realize that our decision will limit to varying degrees GAO’s ability to study any industry with firms that make several products or have several buyers.