delivered the opinion of the Court.

The Miller Act, 49 Stat. 793, as amended, 80 Stat. 1139, 40 U. S. C. § 270a et seq., requires a Government contractor1 to post a surety bond “for the protection of all persons supplying labor and material in the prosecution of the work provided for” in the contract. The Act further provides that any person who has so furnished labor or material and who has not been paid in full within 90 days after the last labor was performed or material supplied may bring suit on the payment bond for the unpaid balance. 40 U. S. C. § 270b (a). This case presents several unresolved issues of importance in the administration of the Act.

I

Between 1961 and 1968, petitioner F. D. Rich Co. was the prime contractor on numerous federal housing projects. During the years 1963-1966, much if not all of the plywood and millwork for these projects was supplied by Cerpac Co. The Cerpac organization was closely intertwined with Rich. The principals of Rich held a substantial voting interest in Cerpac stock, supplied a major share of its working capital, and were thoroughly familiar with its operations and financial condition.

On October 18, 1965, Rich contracted with the United States to build 337 family housing units at Beale Air *119Force Base, California. Rich’s Miller Act surety, petitioner Transamerica Insurance Co., posted the payment bond required by the Act. Rich then awarded Cerpac two contracts on the project, one to select, modify, detail, and install all custom millwork, and one merely to supply all standard exterior plywood, each contract incorporating by reference terms of the prime contract. A similar arrangement was employed by Rich and Cerpac on other projects during this period.

On February 22, 1966, Cerpac placed a single order with respondent Industrial Lumber Co. for all exterior plywood required under its plywood contract for the Beale project. Industrial is a broker, purchasing wood products and materials for resale. It acknowledged the complete Cerpac order, purchased the plywood from its own suppliers and arranged for deliveries at the Beale site to begin on March 10, 1966. Each shipment was receipted as it arrived on the site by a Rich representative.

Shortly after Industrial’s shipments began, Rich informed Cerpac that more plywood was needed for another Government project being constructed in Charleston, South Carolina, for which Cerpac had also contracted to supply Rich with all exterior plywood. Rich and Cer-pac decided to divert some of the Beale lumber to Charleston. Accordingly, Industrial was advised to supply, a shipment of the plywood called for under its Beale contract with Cerpac to the South Carolina site. Industrial arranged for the wood to be shipped by one of its suppliers to a railhead near Charleston. The shipment diverted to South Carolina was one of 22 called for by Industrial’s Beale Contract.2 There were several subsequent shipments to the California site under that contract.

*120During April and May 1966, Cerpac fell behind in its payments to Industrial, and on July 13, 1966, having not received payment on invoices for nine separate shipments, Industrial gave notice to Rich and its surety of a Miller Act claim and thereafter brought the instant action in the Federal District Court for the Eastern District of California.3 The District Court recognized that under our decision in MacEvoy Co. v. United States ex rel. Tomkins Co., 322 U. S. 102 (1944), Rich’s liability turned on whether Cerpac was a “subcontractor” within the meaning of the Act or merely a materialman. The District Court found that Cerpac was a subcontractor; hence Industrial, as its supplier, could assert a Miller Act claim against Rich, the prime contractor on the project. The District Court also rejected Rich’s claim that venue for suit on the South Carolina shipment was improper in the Eastern District of California. Accordingly, the District Court granted judgment for Industrial, holding Cerpac4 and Rich as primary obligees and Transamerica on its bond, jointly and severally liable for the amount of all nine unpaid invoices, $31,402.97, including the amount *121due on the shipment diverted to South Carolina. The District Court, however, denied Industrial’s claim for attorneys’ fees.

Both Rich and Industrial appealed. The Court of Appeals affirmed the judgment against Rich in large part.5 On Industrial’s cross-appeal, the court reversed, holding that attorneys’ fees should have been awarded to Industrial as a successful plaintiff under the Miller Act, and remanded to the District Court for consideration of the amount of attorneys’ fees to be awarded. 473 F. 2d 720 (CA9 1973). We granted certiorari.6 414 U. S. 816 (1973). We affirm the judgment below to the extent it holds that Cerpac was a “subcontractor” for Miller Act purposes and that there was proper venue, but reverse as to the propriety of an award of attorneys’ fees.

II

Section 270a (a) (2) of the Miller Act establishes the general requirement of a payment bond to protect those who supply labor or materials to a contractor on a federal *122project. Ordinarily, a supplier of labor or materials on a private construction project can secure a mechanic’s lien against the improved property under state law. But a lien cannot attach to Government property, see Illinois Surety Co. v. John Davis Co., 244 U. S. 376, 380 (1917), so suppliers on Government projects are deprived of their usual security interest. The Miller Act was intended to provide an alternative remedy to protect the rights of these suppliers.

The rights afforded by the Act are limited, however, by the proviso of § 270b (a). In MacEvoy Co. v. United States ex rel. Tomkins Co., supra, this Court construed § 270b (a) to limit the protection of a Miller Act bond to those who had a contractual agreement with the prime contractor or with a “subcontractor.” Those in more remote relationships, including persons supplying labor or material to a mere materialman, were found not to be protected. 322 U. S., at 109-111. Industrial was a supplier of materials to Cerpac. Thus, if Cerpac were a subcontractor for purposes of the Act, Industrial, having given the required statutory notice, could assert a Miller Act claim against Rich, the prime contractor. But, if Cerpac were merely a materialman, Industrial could not assert its Miller Act claim since it would be merely a supplier of materials to a materialman, a relationship found too remote in MacEvoy to enjoy the protections of the Act.

Petitioners assert that the courts below erred in finding Cerpac a subcontractor. Cerpac’s role under the plywood contract alone was that of a broker receiving standard lumber supplied by Industrial and, in turn, supplying it without modification to Rich. Petitioners argue that the court should not have looked beyond the plywood contract to determine Cerpac’s status under the Act.

*123In MacEvoy, supra, the Court adopted a functional rather than a technical definition for the term subcontractor, as used in the proviso. The Court noted that a subcontractor is “one who performs for and takes from the prime contractor a specific part of the labor or material requirements of the original contract . . . 322 U. S., at 109. The Court went on to explain the reason for the exclusion from the protections of the Act of suppliers of mere materialmen as opposed to those who su]D-ply subcontractors:

“The relatively few subcontractors who perform part of the original contract represent in a sense the prime contractor and are well known to him. It is easy for the prime contractor to secure himself against loss by requiring the subcontractors to give security by bond, or otherwise, for the payment of those who contract directly with the subcontractors. . . . But this method of protection is generally inadequate to cope with remote and undeterminable liabilities incurred by an ordinary materialman, who may be a manufacturer, a wholesaler or-a retailer.” Id., at ,110. (Emphasis added.)

The Court of Appeals properly construed our holding in MacEvoy to establish as a test for whether one is a subcontractor, the substantiality and importance of his relationship with the prime contractor.7 It is the sub-stantiality of the relationship which will usually determine whether the prime contractor can protect himself, since he can easily require bond security or other protection from those few “subcontractors” with whom he has a *124substantial relationship in the performance of the contract.

Measured against that test, Cerpac was clearly a subcontractor for the purposes of the Act. The Miller Act is “highly remedial [and] entitled to a liberal construction and application in order properly to effectuate the Congressional intent to protect those whose labor and materials go into public projects.” MacEvoy, supra, at 107. It is consistent with that intent to look at the total relationship between Cerpac and Rich, not just the contract to supply exterior plywood, to determine whether Cerpac was a subcontractor.8 Cerpac had not only agreed to supply standard plywood but also had a separate contract to select, modify, detail, and install all custom mill-work for the Beale project. Cerpac, in effect, took over a substantial part of the prime contract itself. Moreover, the management and financial structures of the two companies were closely interrelated and their relationship on the Beale project was the same as on many other similar Government projects during the same period. Cerpac was, as the Court of Appeals observed, “in a special, integral, almost symbiotic relationship [with] Rich.” 473 F. 2d, at 724. It would have been easy for Rich to secure itself from loss as a result of a default by Cerpac.

Ill

We also agree with the courts below that venue under the Miller Act for suit on the shipment diverted to South Carolina properly lay in the Eastern District of California. The Act provides:

“Every suit instituted under this section shall be brought in . . . the United States District Court for *125any district in which the contract was to be performed and executed and not elsewhere . . . .” 40 U.S.C. §270b (b).

Petitioners argue that this provision bars a district court in California9 from adjudicating respondent’s claims arising from the shipments of plywood delivered in South Carolina. But § 270b (b) is merely a venue requirement10 and there was clearly a sufficient nexus for its satisfaction. The “Beale 647” contract between Cerpac and Industrial was executed in California, all of the materials described therein to be delivered to a worksite in that State. Although one of the 22 shipments made pursuant to the contract was later diverted to South Carolina for petitioner Rich’s convenience, the site for performance of the original contract remained the same for Miller Act purposes.11 Several shipments to the Beale site were made after the South Carolina shipment. Moreover, petitioners have pointed to no prejudice resulting from the case’s being heard in the California court and considerations of judicial economy and convenience clearly support venue in the *126District Court where all of respondent’s claims arising from the “Beale 647” contract could be adjudicated in a single proceeding.

IV

We turn now to the question of whether attorneys’ fees were properly awarded respondent as a successful Miller Act plaintiff. The so-called “American Rule” governing the award of attorneys’ fees in litigation in the federal courts is that attorneys’ fees “are not ordinarily recoverable in the absence of a statute or enforceable contract providing therefor.” Fleischmann Distilling Corp. v. Maier Brewing Co., 386 U. S. 714, 717 (1967). There was no contractual provision concerning attorneys’ fees in this case. Nor does the Miller Act explicitly provide for an award of attorneys’ fees to a successful plaintiff. But the Court of Appeals construed the Act to require an award of attorneys’ fees where the “public policy” of the State in which suit was brought allows for the award of fees in similar contexts. The court reasoned that the Act provides remedies “ ‘in lieu of the lien upon land and buildings customary where property is owned by private persons’ .... The federal remedy was intended to substitute for the unavailable state remedy of the lien. Therefore, if state [law] allows a supplier on private projects to recover such fees, there is no reason for a different rule to apply to federal projects . . . .”12 Looking to California law, the Court of Appeals found an award of attorneys’ fees proper because Cal. Govt. Code § 4207 (1966) allowed for the recovery of at*127torneys' fees in state actions on the bonds of contractors for state and municipal public works projects.13

We think the Court of Appeals erred in its construction of the statute. The Miller Act provides a federal cause of action, and the scope of the remedy as well as the substance of the rights created thereby is a matter of federal not state law. Neither respondent nor the court below offers any evidence of congressional intent to incorporate state law to govern such an important element of Miller Act litigation as liability for attorneys’ fees. Many federal contracts involve construction in more than one State, and often, as here, the parties to Miller Act litigation have little or no contact, other than the contract itself, with the State in which the federal project is located. The reasonable expectations of such potential litigants are better served by a rule of uniform national application.

A uniform rule also avoids many of the pitfalls which have already manifested themselves in using state law referents. For example, California law does not provide for awards of attorneys’ fees in suits arising from private construction projects. And, a California court had held that the state statute providing for awards of attorneys’ fees in suits on the bonds of state and municipal.public works contractors is inapplicable to construction projects of the United States.14 The Court of Appeals nonethe*128less held that since federal law controls Miller Act recoveries, it was free to look to “state policy” rather than state law and proceeded to find an award of attorneys' fees appropriate. Although the court below premised its decision on the theory that a Miller Act remedy is afforded “ 'in lieu of the lien upon land and buildings customary where property is owned by private persons,’ ” it gave respondent more protection than California law affords litigants involved in disputes arising from private construction projects who are not entitled to an award of attorneys’ fees. We think it better to extricate the federal courts from the morass of trying to divine a “state policy” as to the award of attorneys’ fees in suits on construction bonds.

Finally, the Court of Appeals intimates that in providing that Miller Act claimants should recover the “sums justly due,” 40 TJ. S. C. § 270b (a), Congress must have intended to provide for the award of attorneys’ fees because without such fee shifting, Miller Act claimants would not be fully compensated — the claimant’s recovery would always be diminished by the cost of his legal representation. This argument merely restates one of the oft-repeated criticisms of the American Rule.15 Almost a half century ago, the Massachusetts Judicial Council pleaded for reform, asking, “On what principle of justice can a plaintiff wrongfully run down on a public highway recover *129his doctor’s bill but not his lawyer’s bill?” 16 We recognize that there is some force to the argument that a party who must bear the costs of his attorneys’ fees out of his recovery is not made whole. But there are countervailing considerations as well. We have observed that “one should not be penalized for merely defending or prosecuting a lawsuit, and that the poor might be unjustly discouraged from instituting actions to vindicate their rights if the penalty for losing included the fees of their opponents’ counsel.” Fleischmann Distilling Corp. v. Maier Brewing Co., 386 U. S. 714, 718 (1967). Moreover, “the time, expense, and difficulties of proof inherent in litigating the question of what constitutes reasonable attorney’s fees,” ibid., has given us pause, even though courts have regularly engaged in that endeavor in the many contexts where fee shifting is mandated by statute, policy, or contract. Finally, there is the possibility of a threat being posed to the principle of independent advocacy by having the earnings of the attorney flow from the pen of the judge before whom he argues.

The American Rule has not served, however, as an absolute bar to the shifting of attorneys’ fees even in the absence of statute or contract. The federal judiciary has recognized several exceptions to the general principle that each party should bear the costs of its own legal representation. We have long recognized that attorneys’ fees may be awarded to a successful party when his opponent has acted in bad faith, vexatiously, wantonly, or for oppressive reasons,17 or where a successful litigant *130has conferred a substantial benefit on a class of persons and the court’s shifting of fees operates to spread the cost proportionately among the members of the benefited class.18 The lower courts have also applied a rationale for fee shifting based on the premise that the expense of litigation may often be a formidable if not insurmountable obstacle to the private litigation necessary to enforce important public policies.19 This “private attorney general” rationale has not been squarely before this Court and it is not so now; nor do we intend to imply any view either on the validity or scope of that doctrine. It is sufficient for our purposes here to observe that this case clearly does not fall within any of these exceptions.

Miller Act suits are plain and simple commercial litigation. In effect then, we are being asked to go the last mile in this case, to judicially obviate the American Rule in the context of everyday commercial litigation, where *131the policies which underlie the limited judicially created departures from the rule are inapplicable. This we are unprepared to do. The perspectives of the profession, the consumers of legal services, and other interested groups should be weighed in any decision to substantially undercut the application of the American Rule in such litigation. Congress is aware of the issue.20 Thus whatever the merit of arguments for a further departure from the American Rule in Miller Act commercial litigation, those arguments are properly addressed to Congress.

The judgment of the Court of Appeals is affirmed insofar as it holds that Cerpac is a subcontractor for Miller Act purposes and that there was proper venue for suit on the shipment diverted to South Carolina, but reversed insofar as it holds that an award of attorneys’ fees to respondent Industrial is required by the Act.

It is so ordered.

dissenting in part.

The Court, dealing with the Miller Act’s predecessor, held in Illinois Surety Co. v. John Davis Co., 244 U. S. 376, 380, that the Heard Act “must be construed liberally.” That same principle applies to the Miller Act. Fleisher Co. v. United States ex rel. Hallenbeck, 311 U. S. 15, 17-18. The Act is silent as to attorneys’ fees, saying only that the payment bond shall allow the supplier “to prosecute said action to final execution and judgment for the sum or sums justly due him.” 40 U. S. C. § 270b (a).

*132The Miller Act is unlike the Lanham Act involved in Fleischmann Distilling Corp. v. Maier Brewing Co., 386 U. S. 714. That Act itemized the components of the remedy which the Act afforded: injunctive relief, treble damages, and “costs” (which by federal statute did not include attorneys’fees). Id., at 719-720. Moreover, attempts to amend the Lanham Act to include attorneys’ fees had never succeeded, id., at 721. Here there is no such legislative history; nor does the Miller Act itemize the components of the “sum or sums justly due.”

The Court says that dependence on state law is inappropriate, for we deal with a federal standard that should be uniform. That takes great liberties with the Miller Act. Here the contract and law were made in California and were to be performed there. In Illinois Surety Co. v. John Davis Co., supra, the contract and law were made in Illinois and were to be performed there. “Questions of liability for interest must therefore be determined by the law of that State,” said Mr. Justice Brandéis speaking for the Court, 244 U. S., at 381. If state law would give the claimant interest, it should give him attorneys’ fees based on the purpose of the Miller Act. Judge Carter writing for the Court of Appeals pointed out that the Miller Act is the federal equivalent of state lien laws. See 473 F. 2d 720, 727. The remedy in a federal suit is therefore properly composed of the same elements as would be available to lien claimants in a state court collecting for labor and materials furnished on nonfederal projects. One of the elements of recovery permitted in a California court is attorneys' fees. The “sum or sums justly due” should as a matter of federal law be construed to be the same as that -due a claimant whose remedy is based on a state statute, when the federal remedy was intended to be the equivalent of the state remedy.

*133What Mr. Justice Brandéis said of interest is equally applicable to attorneys’ fees under the Miller Act.* Under the circumstances present here it would seem quite unjust not to include in the sum that is due the cost of collecting that sum.