UNITED STATES v. NEW MEXICO

JUSTICE BLACKMUN delivered the opinion of the Court.

We are presented here with a recurring problem: to what extent may a State impose taxes on contractors that conduct business with the Federal Government?

I

A

This case concerns the contractual relationships between three private entities and the United States. The three agreements involved are typical in most respects of management contracts devised by the Atomic Energy Commission (AEC), now the Department of Energy (DOE).1 Like many of the Government's contractual undertakings, DOE management contracts generally provide the private contractor with its costs plus a fixed fee. But in several ways DOE agreements are a unique species of contract, designed to facilitate long-term private management of Government-owned research and development facilities. As the parties to this case acknowledge, the complex and intricate contractual provisions make it virtually impossible to describe the contractual relationship in standard agency terms. See App. 196-197; Hiestand & Florsheim, The AEC Management Contract Concept, 29 Federal B. J. 67 (1969) (Hiestand & Florsheim). While subject to the general direction of the Government, the contractors are vested with substantial autonomy in their operations and procurement practices.2

The first of the contractors, Sandia Corporation, was organized in 1949 as a subsidiary of Western Electric Company, Inc. Sandia manages the Government-owned Sandia Laboratories in Albuquerque, N. M., and engages exclusively in federally sponsored research. It receives no fee under its contract, and owns no property except for $1,000 in United States bonds that constitute its paid-in capital. But Sandia and Western Electric are guaranteed royalty-free, irrevocable licenses for any communications-related discoveries or inventions developed by most Sandia employees during the course of the contract, App. 34-35, and the company receives complete reimbursements for salary outlays and other expenditures. Id., at 40-42.3

The Zia Company, another of the contractors, is a subsidiary of Santa Fe Industries, Inc. Since 1946, Zia has performed a variety of management, maintenance, and related functions at the Government's Los Alamos Scientific Laboratory, for which it receives its costs as well as a fixed annual fee. While Zia owns property and performs private work, virtually none of its property is used in the performance of its contract with the Government, and all of its private activities are conducted away from Los Alamos by a separate work force.

The third contractor is Los Alamos Constructors, Inc. (LACI), since 1953 a subsidiary of Zia. LACI's operations are limited to construction and repair work at the Los Alamos facility. The company owns no tangible personal property and makes no purchases; it procures needed property and equipment through its parent, Zia. And like Zia, LACI receives its costs plus a fixed annual fee from the Government.

The management contracts between the Government and the three contractors have a number of significant features in common. As in most DOE atomic facility management agreements, the contracts provide that title to all tangible personal property purchased by the contractors passes directly from the vendor to the Government. App. 231a (Zia); id., at 34 (Sandia).4 Similarly, the Government bears the risk of loss for property procured by the contractors. Zia and LACI must submit an annual voucher of expenditures for Government approval. Id., at 20 (Zia); id., at 27 (LACI).5 And the agreements give the Government control over the disposition of all property purchased under the contracts, as well as over each contractor's property management procedures. Disputes under the contracts are to be resolved by a DOE contracting official. Id., at 128-129 (Zia standard terms) and 157-158 (Sandia standard terms).

On the other hand, the contractors place orders with third-party suppliers in their own names, and identify themselves as the buyers. See id., at 36-37 (Sandia contract) and 120 (Zia standard terms). Indeed, the Government acknowledged during discovery that Sandia, Zia, and LACI "may be. . . `independent contractor[s],' rather than . . . `servant[s]' for . . . given `function[s] under' the contract[s] (e. g., directing the details of day-to-day . . . operations and the hiring and direct supervision of employees)," id., at 197, and the Government does not claim that the contractors are federal instrumentalities. Id., at 201; see Department of Employment v. United States, 385 U.S. 355 (1966). Similarly, the United States disclaims responsibility for torts committed by the contractors' employees, and maintains that such employees have no claim against the United States for labor-related grievances. See 624 F.2d 111, 116-117, n. 6 (CA10 1980).

Finally, and most importantly, the contracts use a so-called "advanced funding" procedure to meet contractor costs. Advanced funding, an accounting device developed shortly after the conclusion of the Manhattan Project, is designed to provide "up-to-date meaningful records of costs and controls of property," as well as to "speed up reimbursement of contractors." App. 204 (Fifth Semiannual Report of the Atomic Energy Commission (1949)). The procedure allows contractors to pay creditors and employees with drafts drawn on a special bank account in which United States Treasury funds are deposited.6

To put the advanced funding mechanism in place, the United States, the contractor, and a bank establish a designated bank account, pursuant to a three-party contract. The Government dispatches a letter of credit to a Federal Reserve Bank in favor of the contractor, making Treasury funds available in the designated account. The contractor pays its expenses by drawing on the account, at which time the bank or the contractor executes a payment voucher in an amount sufficient to cover the draft. The voucher is forwarded to the Federal Reserve Bank. The United States owns the account balance. See id., at 19-20, 84-90a, 109-113. As a result of all this, only federal funds are expended when the contractor makes purchases. If the Government fails to provide funding, the contractor is excused from performance of the contract, and the Government is liable for all properly incurred claims.

Prior to July 1, 1977, the Government's contracts with Sandia, Zia, and LACI did not refer to the contractors as federal "agents." On that date — some two years after the commencement of this litigation — the agreements were modified to state that each contractor "acts as an agent [of the Government]. . . for certain purposes," including the disbursement of Government funds and the "purchase, lease, or other acquisition" of property. Id., at 50-51, 55-56, 59-60. This was designed to recognize what was described as the "longstanding agency status and authority" of the contractors. Id., at 50, 55, 59. Thus it was made clear that Sandia and Zia were authorized to "pledge the credit of the United States," id., at 52 and 56, and the Government declared that it "considers all obligations properly incurred" in accordance with the contractual provisions to be Government obligations "from their inception." Id., at 52 (Sandia), 56 (Zia), and 60 (LACI). At the same time, however, the United States denied any intent "formally and directly [to] designat[e] the contractors as agents," id., at 64, and each modification stated that it did not "create rights or obligations not otherwise provided for in the contract." Id., at 52, 57, 61.

B

New Mexico imposes a gross receipts tax and a compensating use tax on those doing business within the State. With limited exceptions, "[f]or the privilege of engaging in business, an excise tax equal to four per cent [4%] of gross receipts is imposed on any person engaging in business in New Mexico." N. M. Stat. Ann. § 72-16A-4 (Supp. 1975).7 In effect, the gross receipts tax operates as a tax on the sale of goods and services. The State also levies a compensating use tax, equivalent in amount to the gross receipts tax, "[f]or the privilege of using property in New Mexico." § 72-16A-7. This is imposed on property acquired out-of-state in a "transaction that would have been subject to the gross receipts tax had it occurred within [New Mexico]." § 72-16A-7(A)(2).8 Thus the compensating use tax functions as an enforcement mechanism for the gross receipts tax by imposing a levy on the use of all property that has not already been taxed; the State collects the same percentage regardless of where the property is purchased. Neither tax, however, is imposed on the "receipts of the United States or any agency or instrumentality thereof," or on the "use of property by the United States or any agency or instrumentality thereof." §§ 72-16A-12.1, 72-16A-12.2.

Without objection, Zia and LACI each year paid the New Mexico gross receipts tax on the fixed fees they received from the Federal Government. But the Government argued that the contractors' other expenditures and operations are constitutionally immune from state taxation. In July 1975 the United States therefore initiated this suit in the United States District Court for the District of New Mexico, seeking a declaratory judgment that advanced funds are not taxable gross receipts to the contractors; that the receipts of vendors selling tangible property to the United States through the contractors cannot be taxed by the State; and that the use of Government-owned property by the contractors is not subject to the State's compensating use tax. App. 11-12.9

The District Court granted the United States summary judgment. Relying on Kern-Limerick, Inc. v. Scurlock, 347 U.S. 110 (1954), the court determined that the crucial inquiry is whether the contractors are "procurement agents" for the Government. The court answered that question in the affirmative, noting that the Government "maintains control over the contractors' procurement systems, property management and disposal practices, method of payment of operational costs, and other operations under the contracts." 455 F.Supp. 993, 997 (1978). That analysis led the court to identify an agency relationship existing even in the years prior to the 1977 contract modifications. Ibid. The court therefore held that the gross receipts tax cannot constitutionally be applied to purchases by the contractors; because the court viewed the compensating use tax as a correlative of the receipts tax, it determined that the use tax also was invalid as applied to Sandia, Zia, and LACI. Id., at 998. Finally, the court ruled that advanced funds do not serve as compensation to the contractors, and therefore cannot be taxed as gross receipts. Id., at 998-999.

The United States Court of Appeals for the Tenth Circuit reversed. 624 F.2d 111 (1980). In its view, this Court's decisions in the tax immunity area have been "more concerned with preserving the delicate financial balance between our co-existing sovereignties than with rigid adherence to agency law terminology." Id., at 116. Advanced funding, the court declared, "is simply another means of reimbursement devised by accountants to eliminate major weaknesses in the government's bookkeeping practices." Id., at 119. In meeting overhead and salaries with Government funds, the contractors were satisfying their own obligations, and they exercised dominion over the funds by issuing drafts to obligees. And insofar as the claims of third-party vendors are concerned, the court found federal "responsibility for properly incurred claims to be inherent in all cost-type contracts," id., at 119, n. 9; any number of businesses act under letters of credit from banks and other sureties, and the Federal Government itself finances a variety of organizations — incleding States and local governments — in such a manner.

The other contractual provisions relied on by the District Court — federal control over procurement systems, management practices, and the like — failed to impress the Court of Appeals. It concluded that the Government-contractor relationship, viewed as a whole, did not "`so incorporat[e] [the contractors] into the government structure as to [make them] instrumentalities of the United States . . . .' " Id., at 118, quoting United States v. Boyd, 378 U.S. 39, 48 (1964). And that Sandia received no fee for its services was of little consequence, in the court's view, because "decisions on the amount of fee, if any, to be paid a government contractor are not made primarily with agency consequences in mind." 624 F. 2d, at 120. Since the 1977 contractual amendments by their terms added nothing of substance to the agreements, they did not affect the court's analysis. The District Court was directed to enter summary judgment for New Mexico. Id., at 121.

The United States sought certiorari, and we granted the writ to consider the seemingly intractable problems posed by state taxation of federal contractors. 450 U.S. 909 (1981).

II

A

With the famous declaration that "the power to tax involves the power to destroy," McCulloch v. Maryland, 4 Wheat. 316, 431 (1819), Chief Justice Marshall announced for the Court the doctrine of federal immunity from state taxation. In so doing he introduced the Court to what has become a "much litigated and often confused field," United States v. City of Detroit, 355 U.S. 466, 473 (1958), one that has been marked from the beginning by inconsistent decisions and excessively delicate distinctions.

McCulloch itself relied on generalized notions of federal supremacy to invalidate a state tax on the Second Bank of the United States. The Court gave broad scope to state power: the opinion declined to "deprive the States of any resources which they originally possessed. It does not extend to . . . a tax imposed on the interest which the citizens of Maryland may hold in [the Bank], in common with other property of the same description throughout the State." 4 Wheat., at 436. Not long afterwards, however, Chief Justice Marshall, speaking for the Court, seemingly disregarded the McCulloch dictum in striking down a state tax on interest income from federal bonds, explaining that such levies cannot constitutionally fall on an "operation essential to the important objects for which the government was created." Weston v. City Council of Charleston, 2 Pet. 449, 467 (1829). During the following century the Court took to heart Weston's expansive analysis of federal tax immunity, invalidating, among many others, state taxes on the income of federal employees, Dobbins v. Commissioners of Erie County, 16 Pet. 435 (1842); on income derived from property leased from the Federal Government, Gillespie v. Oklahoma, 257 U.S. 501 (1922); and on sales to the United States, Panhandle Oil Co. v. Mississippi ex rel. Knox, 277 U.S. 218 (1928).10

These decisions, it has been said, were increasingly divorced both from the constitutional foundations of the immunity doctrine and from "the actual workings of our federalism," Graves v. New York ex rel. O'Keefe, 306 U.S. 466, 490 (1939) (Frankfurter, J., concurring), and in James v. Dravo Contracting Co., 302 U.S. 134 (1937), by a 5-4 vote, the Court marked a major change in course. Over the dissent's justifiable objections that it was "overrul[ing], sub silentio, a century of precedents," id., at 161, the Court upheld a state tax on the gross receipts of a contractor providing services to the Federal Government:

The Court's more recent cases involving federal contractors generally have hewed to the James analysis. Alabama v. King & Boozer, 314 U.S. 1 (1941), upheld a state tax on sales to a federal contractor, overruling Panhandle Oil Co. v. Mississippi ex rel. Knox, supra. Decisions such as United States v. City of Detroit, supra, have validated state use taxes on private entities holding federal property.

Even the Court's post-James decisions, however, cannot be set in an entirely unwavering line. United States v. Allegheny County, 322 U.S. 174 (1944), invalidated a state property tax that included in the assessment the value of federal machinery held by a private party; 14 years later that decision in large part was overruled by United States v. City of Detroit, supra. See United States v. County of Fresno, 429 U.S. 452, 462-463, n. 10 (1977). In Livingston v. United States, 364 U.S. 281 (1960), summarily aff'g 179 F.Supp. 9 (EDSC 1959), the Court, without opinion or citation, approved the invalidation of a state use tax as applied to a federal contractor. Yet United States v. Boyd, supra, upheld a virtually identical state tax, seemingly confining Livingston to its "extraordinary" facts. 378 U. S., at 45, n. 6.

Similarly, the decisions fail to speak with one voice on the relevance of traditional agency rules in determining the tax-immunity status of federal contractors. Thus, Alabama v. King & Boozer, supra, declined to find immunity in part because the contractors involved lacked the "status of agents," 314 U. S., at 13, and United States v. Township of Muskegon, 355 U.S. 484, 486 (1958), upheld a use tax on a federal contractor with the caveat that the "case might well be different if [the contractor] . . . could properly be called a `servant' of the United States in agency terms." See Kern-Limerick, Inc. v. Scurlock, 347 U.S. 110 (1954). Yet James v. Dravo Contracting Co., supra, stated flatly that tax immunity is not dependent "`upon the nature of the agents, or upon the mode of their constitution, or upon the fact that they are agents.' " 302 U. S., at 154, quoting Railroad Co. v. Peniston, 18 Wall. 5, 36 (1873) (plurality opinion). And United States v. Boyd, supra, rejected the Government's argument that its contractors were federal agents and therefore tax immune, stating simply that the private entities were not "instrumentalities of the United States." 378 U. S., at 48.

B

We have concluded that the confusing nature of our precedents counsels a return to the underlying constitutional principle. The one constant here, of course, is simple enough to express: a State may not, consistent with the Supremacy Clause, U. S. Const., Art. VI, cl. 2, lay a tax "directly upon the United States." Mayo v. United States, 319 U.S. 441, 447 (1943). While "[o]ne could, and perhaps should, read M'Culloch . . . simply for the principle that the Constitution prohibits a State from taxing discriminatorily a federally established instrumentality," First Agricultural Bank v. State Tax Comm'n, 392 U.S. 339, 350 (1968) (dissenting opinion), the Court has never questioned the propriety of absolute federal immunity from state taxation. And after 160 years, the doctrine has gathered "a momentum of authority that reflects, if not a detailed exposition of considerations of policy demanded by our federal system, certainly a deep instinct that there are such considerations . . . ." City of Detroit v. Murray Corp., 355 U.S. 489, 503-504 (1958) (opinion of Frankfurter, J.).

But the limits on the immunity doctrine are, for present purposes, as significant as the rule itself. Thus, immunity may not be conferred simply because the tax has an effect on the United States, or even because the Federal Government shoulders the entire economic burden of the levy. That is the import of Alabama v. King & Boozer, where a sales tax was imposed on the gross receipts of a vendor selling to a cost-plus Government contractor. The Court found it constitutionally irrelevant that the United States reimbursed all the contractor's expenditures, including those going to meet the tax: the Government's right to be free from state taxation "does not spell immunity from paying the added costs, attributable to the taxation of those who furnish supplies to the Government and who have been granted no tax immunity." 314 U. S., at 9. That the contractor is purchasing property for the Government is similarly irrelevant; in King & Boozer, title to goods purchased by the contractor vested in the United States immediately upon shipment by the seller. Id., at 13.

Similarly, immunity cannot be conferred simply because the state tax falls on the earnings of a contractor providing services to the Government. James v. Dravo Contracting Co., supra. And where a use tax is involved, immunity cannot be conferred simply because the State is levying the tax on the use of federal property in private hands, United States v. City of Detroit, supra, even if the private entity is using the Government property to provide the United States with goods, United States v. Township of Muskegon, supra; City of Detroit v. Murray Corp., supra, or services, Curry v. United States, 314 U.S. 14 (1941); United States v. Boyd, supra. In such a situation the contractor's use of the property "in connection with commercial activities carried on for profit," is "a separate and distinct taxable activity." United States v. Boyd, 378 U. S., at 44. Indeed, immunity cannot be conferred simply because the tax is paid with Government funds; that was apparently the case in Boyd, where the contractor made expenditures under an advanced funding arrangement similar to the one involved here. Id., at 41.

What the Court's cases leave room for, then, is the conclusion that tax immunity is appropriate in only one circumstance: when the levy falls on the United States itself, or on an agency or instrumentality so closely connected to the Government that the two cannot realistically be viewed as separate entities, at least insofar as the activity being taxed is concerned. This view, we believe, comports with the principal purpose of the immunity doctrine, that of forestalling "clashing sovereignty," McCulloch v. Maryland, 4 Wheat., at 430, by preventing the States from laying demands directly on the Federal Government. See City of Detroit v. Murray Corp., 355 U. S., at 504-505 (opinion of Frankfurter, J.). As the federal structure — along with the workings of the tax immunity doctrine11 — has evolved, this command has taken on essentially symbolic importance, as the visible "consequence of that [federal] supremacy which the constitution has declared." McCulloch v. Maryland, 4 Wheat., at 436. At the same time, a narrow approach to governmental tax immunity accords with competing constitutional imperatives, by giving full range to each sovereign's taxing authority. See Graves v. New York ex rel. O'Keefe, 306 U. S., at 483.

Thus, a finding of constitutional tax immunity requires something more than the invocation of traditional agency notions: to resist the State's taxing power, a private taxpayer must actually "stand in the Government's shoes." City of Detroit v. Murray Corp., 355 U. S., at 503 (opinion of Frankfurter, J.). That conclusion is compelled by the Court's principal decisions exploring the nature of the Constitution's immunity guarantee. Chief Justice Hughes' opinion for the Court in James, which set the doctrine on its modern course, suggested that a state tax is impermissible when the taxed entity is "so intimately connected with the exercise of a power or the performance of a duty" by the Government that taxation of it would be " `a direct interference with the functions of government itself.' " 302 U. S., at 157, quoting Metcalf & Eddy v. Mitchell, 269 U.S. 514, 524 (1926). And the point is settled by Boyd, the Court's most recent decision in the field. There, the Government argued that its contractors were tax-exempt because they were federal agents. Without any discussion of traditional agency rules the Court rejected that suggestion out-of-hand, declaring that "we cannot believe that [the contractors are] `so assimilated by the Government as to become one of its constituent parts.' " 378 U. S., at 47, quoting United States v. Township of Muskegon, 355 U. S., at 486. And the Court continued:

The Court's other cases describing the nature of a federal instrumentality have used similar language: "virtually . . . an arm of the Government," Department of Employment v. United States, 385 U. S., at 359-360; "integral parts of [a governmental department]," and "arms of the Government deemed by it essential for the performance of governmental functions," Standard Oil Co. v. Johnson, 316 U.S. 481, 485 (1942).

Granting tax immunity only to entities that have been "incorporated into the government structure" can forestall, at least to a degree, some of the manipulation and wooden formalism that occasionally have marked tax litigation — and that have no proper place in determining the allocation of power between coexisting sovereignties. In this case, for example, the Government and its contractors modified their agreements two years into the litigation in an obvious attempt to strengthen the case for nonliability. Yet the Government resists using its own employees for the tasks at hand — or, indeed, even formally designating Sandia, Zia, and LACI as agents — because it seeks to tap the expertise of industry, without subjecting its contractors to burdensome federal procurement regulations. See Hiestand & Florsheim, at 81; App. 182-184. Instead, the Government earnestly argues that its contractors are entitled to tax immunity because, among other things, they draw checks directly on federal funds, instead of waiting a time for reimbursement. Brief for United States 32-35. We cannot believe that an immunity of constitutional stature rests on such technical considerations, for that approach allows "any government functionary to draw the constitutional line by changing a few words in a contract." Kern-Limerick, Inc. v. Scurlock, 347 U. S., at 126 (dissenting opinion).

If the immunity of federal contractors is to be expanded beyond its narrow constitutional limits, it is Congress that must take responsibility for the decision, by so expressly providing as respects contracts in a particular form, or contracts under particular programs. James v. Dravo Contracting Co., 302 U. S., at 161; Carson v. Roane-Anderson Co., 342 U.S. 232, 234 (1952). And this allocation of responsibility is wholly appropriate, for the political process is "uniquely adapted to accommodating the competing demands" in this area. Massachusetts v. United States, 435 U.S. 444, 456 (1978) (plurality opinion). See United States v. City of Detroit, 355 U. S., at 474. But absent congressional action, we have emphasized that the States' power to tax can be denied only under "the clearest constitutional mandate." Michelin Tire Corp. v. Wages, 423 U.S. 276, 293 (1976).

III

It remains to apply these principles to the Sandia, Zia, and LACI contracts. The Government concedes that the legal incidence of the gross receipts and use taxes falls on the contractors, Brief for United States 25, and we do not disagree. See United States v. New Mexico, 581 F.2d 803, 806 (CA10 1978). The issue, then, is whether the contractors can realistically be considered entities independent of the United States. If so, a tax on them cannot be viewed as a tax on the United States itself.

So far as the use tax is concerned, United States v. Boyd, supra, controls this case. The contracts at issue in Boyd were standard AEC management contracts, in all relevant respects identical to the ones here. The contractors performed maintenance and construction work at Government facilities, under the general direction of the Government. They procured materials, and paid for the goods with Government funds under an advanced funding arrangement; title passed directly from the vendor to the United States. The contractors owned none of the property involved, and received a fixed annual fee. Indeed, one of the contractor's purchase orders stated that it made purchases "for and on behalf of the Government." 378 U. S., at 42, n. 4. And the Tennessee use tax did not differ in any significant way from the use tax now before us.12

As noted above, the Government argued that this close contractual relationship made the contractors federal agents, and therefore tax immune. Yet the Court had no difficulty upholding the application of the Tennessee tax, concluding that "`[t]he vital thing' is that [the contractors are] `using the property in connection with [their] own commercial activities.' " Id., at 45, quoting United States v. Township of Muskegon, 355 U. S., at 486. That the federal property involved was being used for the Government's benefit — something that by definition will be true in virtually every management contract — was irrelevant, for the contractors remained distinct entities pursuing "private ends," and their actions remained "commercial activities carried on for profit." 378 U. S., at 44. For that reason, the contractors had not become "instrumentalities" of the United States. Id., at 48.13

The same factors are at work here. The tax, the taxed activity, and the contractual relationships do not differ from those involved in Boyd. The contractors here are privately owned corporations; "Government officials do not run [their] day-to-day operations nor does the Government have any ownership interest." First Agricultural Bank v. State Tax Comm'n, 392 U. S., at 354 (dissenting opinion). In contrast to federal employees, then, Sandia and its fellow contractors cannot be termed "constituent parts" of the Federal Government. It is true, of course, that employees are a special type of agent, and like the contractors here employees are paid for their services. But the differences between an employee and one of these contractors are crucial. The congruence of professional interests between the contractors and the Federal Government is not complete; their relationships with the Government have been created for limited and carefully defined purposes. Allowing the States to apply use taxes to such entities does not offend the notion of federal supremacy.14

For similar reasons, the New Mexico gross receipts tax must be upheld as applied to funds received by the contractors to meet salaries and internal costs. Once it is conceded that the contractors are independent taxable entities, it cannot be disputed that their gross income is taxable. This conclusion follows directly from James v. Dravo Contracting Co., supra, where the Court upheld a state tax reaching "gross amounts received from the United States." 302 U. S., at 137. In any event, incurring obligations to achieve contractual ends is not significantly different from using property for the same purposes. And despite the Government's arguments, the use of advanced funding does not change the analysis. That device is, at heart, an efficient method of reimbursing contractors — something the Government has apparently recognized in contexts other than tax litigation. See App. 31 (Sandia contract), 189 (Ninth Semiannual Report of the Atomic Energy Commission (1951)), 191 (same). If receipt of advanced funding is coextensive with status as a federal instrumentality, virtually every federal contractor is, or could easily become, immune from state taxation.

New Mexico's tax on sales to the contractors presents a more complex problem. So far as the use tax discussed above is concerned, the subject of the levy is the taxed entity's beneficial use of the property involved. See United States v. Boyd, 378 U. S., at 44. Unless the entity as a whole is one of the Government's "constituent parts," then, a tax on its use of property should not be seen as falling on the United States; in that situation the property is being used in furtherance of the contractor's essentially independent commercial enterprise. In the case of a sales tax, however, it is arguable that an entity serving as a federal procurement agent can be so closely associated with the Government, and so lack an independent role in the purchase, as to make the sale — in both a real and a symbolic sense — a sale to the United States, even though the purchasing agent has not otherwise been incorporated into the Government structure.

Such was the Court's conclusion in Kern-Limerick, Inc. v. Scurlock, supra, a decision on which the Government heavily relies. The contractor in that case identified itself as a federal procurement agent, and when it made purchases title passed directly to the Government; the purchase orders themselves declared that the purchase was made by the Government and that the United States was liable on the sale. Equally as important, the contractor itself was not liable for the purchase price, and it required specific Government approval for each transaction. See 347 U. S., at 120-121. And, as the Court emphasized, the statutory procurement scheme envisioned the use of federal purchasing agents. Id., at 114. The Court concluded that a sale to the contractor was in effect a sale to the United States, and therefore not a proper subject for the Arkansas sales tax.15 As we have noted elsewhere, Kern-Limerick "stands only for the proposition that the State may not impose a tax the legal incidence of which falls on the Federal Government." United States v. County of Fresno, 429 U. S., at 459-460, n. 7.

We think it evident that the Kern-Limerick principle does not invalidate New Mexico's sales tax as applied to purchases made by the contractors here. Even accepting the Government's representation that it is directly liable to vendors for the purchase price, see Tr. of Oral Arg. 42-45,16 Sandia and Zia nevertheless make purchases in their own names — Sandia, in fact, is contractually obligated to do so, App. 37 — and presumably they are themselves liable to the vendors. Vendors are not informed that the Government is the only party with an independent interest in the purchase, as was true in Kern-Limerick, and the Government disclaims any formal intention to denominate the contractors as purchasing agents. Similarly, Sandia and Zia need not obtain advance Government approval for each purchase.17 These factors demonstrate that the contractors have a substantial independent role in making purchases, and that the identity of interests between the Government and the contractors is far from complete. As a result, sales to Zia and Sandia are in neither a real nor a symbolic sense sales to the "United States itself." It is true that title passes directly from the vendor to the Federal Government, but that factor alone cannot make the transaction a purchase by the United States, so long as the purchasing entity, in its role as a purchaser, is sufficiently distinct from the Government. Alabama v. King & Boozer, 314 U. S., at 13.

There is a final irony in this case. In Carson v. Roane-Anderson Co., 342 U.S. 232 (1952), the Court considered a state sales and use tax imposed on AEC management contractors. The terms of the contracts were in most relevant respects identical to the ones here, and insofar as they differed they established an even closer relationship between the Government and the contractors. See Brief for United States, O. T. 1951, Nos. 186 and 187, pp. 8-12. The Court held that in the last sentence of § 9(b) of the Atomic Energy Act of 1946, 60 Stat. 765 — which barred state or local taxation of AEC "activities" — Congress had statutorily exempted the contractors from state taxation, because the operations of management contractors were Commission activities. 342 U. S., at 234. Congress responded by repealing the last sentence of § 9(b), Pub. L. 262, 67 Stat. 575, in an attempt to "place the Commission and its activities on the same basis, with respect to immunity from State and local taxation, as other Federal agencies." S. Rep. No. 694, 83d Cong., 1st Sess., 3 (1953). In doing so, Congress endorsed the principle that "constitutional immunity does not extend to cost-plus-fixed-fee contractors of the Federal Government, but is limited to taxes imposed directly upon the United States." Id., at 2.

We do not suggest that the repeal of § 9(b) waives the Government's constitutional tax immunity; Congress intended AEC contractors to be shielded by constitutional immunity principles "as interpreted by the courts." S. Rep. No. 694, at 3. But it is worth remarking that DOE is asking us to establish as a constitutional rule something that it was unable to obtain statutorily from Congress. For the reasons set out above, we conclude that the contractors here are not protected by the Constitution's guarantee of federal supremacy. If political or economic considerations suggest that a broader immunity rule is appropriate, "[s]uch complex problems are ones which Congress is best qualified to resolve." United States v. City of Detroit, 355 U. S., at 474.

Accordingly, the judgment of the Court of Appeals is

Affirmed.