concurring.

As I read § 14 (b), the prepositional phrase “in any State or Territory” modifies the immediately preceding noun “employment.” This reading is consistent with the analysis in the Court’s opinion, which I join except for its suggestion that federal policy favors permitting union-shop and agency-shop agreements.

with whom Mr. Justice Rehnquist joins, dissenting.

The respondent, Mobil Oil Corp., is a New York corporation with its home office in New York City. The Gulf-East Coast Operations Division of Mobil’s Marine Transportation Department, located in Beaumont, Tex., operates eight oceangoing American-flag tankers. These ships transport petroleum products between Texas and various ports on the Atlantic coast. Every month each tanker normally makes two round-trip voyages. On the average voyage a ship is at sea for four or five days and spends approximately 18 to 30 hours in port to load or unload its cargo.

The petitioner Maritime Local 8-801 of the Oil, Chemical and Atomic Workers International Union represents the 289 blue-water seamen who man the tankers. When this lawsuit began, 123 of these 289 employees claimed Texas as their residence,1 and 152 of them had requested *423the company to list Beaumont as their home port. Although 40% of the seamen had first applied for work in New York, the remainder had applied for the jobs in Texas, and the final hiring decision for all of them had been made in Beaumont. Texas has collected unemployment compensation insurance premiums for all of these employees, regardless of what State any of them might have designated as his home address or shipping port.

The seamen perform all of their duties aboard ship. They work for approximately 85 days and then receive 37 days of paid shore leave. Eighty to 90% of their work is performed on the high seas.

In 1969 Mobil and the Union concluded a collective-bargaining agreement in New York that covered these seagoing employees. Among other provisions, the agreement contained an agency-shop clause that required all the employees to become “members of the union and/ or in the alternative pay the regular union dues and initiation fees within 31 days from the employment date.” App. 281. In a challenge to this clause, Mobil brought the present suit under § 301 of the Labor Management Relations Act, 1947, 29 U. S. C. § 185, and under 28 U. S. C. § 2201, seeking a declaratory judgment that the clause is invalid because it violates the “right-to-work” laws of Texas.2

The District Court, agreeing with Mobil that Texas was more intimately involved with the employment relationship than any other State, held that Texas' right-to-work laws applied to the agreement. It accordingly declared the agency-shop provision invalid and unenforceable. A divided panel of the Court of Appeals for the Fifth Circuit initially reversed this judgment, 483 F. 2d 603 (1973), but on rehearing en banc, that court af*424firmed the judgment of the District Court. 504 F. 2d 272 (1974). The en banc appellate court analyzed all the relevant contacts that Texas has with the employees in question and concluded that “federal labor legislation, the predominance of Texas contacts over any other jurisdiction, and the significant interest which Texas has in applying its right to work law to this employment relationship warrant application of the Texas law and, consequently, invalidation of the agency shop provision.” Id., at 275.

I

Sections 8 (a) (3) and 14 (b) of the National Labor Relations Act, 29 U. S. C. §§ 158 (a) (3) and 164 (b), delineate the federal interest in union-security arrangements. The first proviso of Section 8(a)(3)3 says *425that no law of the United States “shall preclude an employer from making an agreement with a labor organization ... to require as a condition of employment membership therein on or after the thirtieth day” of employment. The second proviso qualifies the first:

“[N]o employer shall justify any discrimination against an employee for nonmembership in a labor organization (A) if he has reasonable grounds for believing that such membership was not available to the employee on the same terms and conditions generally applicable to other members, or (B) if he has reasonable grounds for believing that membership was denied or terminated for reasons other than the failure of the employee to tender the periodic dues and the initiation fees uniformly required as a condition of acquiring or retaining membership.”

Together, these two provisions sanction a “union shop” agreement, which, although permitting employment of those who are not union members, requires employees to join the union (or pay dues in lieu of membership) 30 days after employment has begun. But they outlaw a “closed shop” agreement, which requires union membership as a precondition to both initial and continued employment. NLRB v. General Motors Corp., 373 U. S. 734, 738-739 (1963).

These provisions modified § 8 (3) of the National Labor Relations Act, 49 Stat. 452, which permitted not only union shops, but closed shops as well. See NLRB v. General Motors Corp., supra, at 739-740; S. Rep. *426No. 105, 80th Cong., 1st Sess., 6 (1947); H. R. Conf. Rep. No. 510, 80th Cong., 1st Sess., 41 (1947). Section 8 (a) (3) was designed to curb the abuses of compulsory unionism, which “create [d] too great a barrier to free employment,” S. Rep. No. 105, supra, at 6, but at the same time, to continue to afford unions a measure of security by enabling them to prevent “free riders.” Id., at 7. As the Court stated in the General Motors case, supra, at 740-741:

“These additions [to § 8 (a)(3)] were intended to accomplish twin purposes. On the one hand, the most serious abuses of compulsory unionism were eliminated by abolishing the closed shop. On the other hand, Congress recognized that in the absence of a union-security provision ‘many employees sharing the benefits of what unions are able to’ accomplish by collective bargaining will refuse to pay their share of the cost.’ ”

Section 8 (a) (3) thus accommodated the competing interests by eliminating the union hiring hall while assuring that “[a]s far as the federal law was concerned, all employees could be required to pay their way.” 373 U. S., at 741; see S. Rep. No. 105, supra, at 6-7.

But Congress chose not to establish a uniform national rule permitting the union shop. States were to be left free to determine that security arrangements of any sort were against the public interest. Algoma Plywood Co. v. Wisconsin Board, 336 U. S. 301, 313-314 (1949). This was made clear in § 14 (b) of the National Labor Relations Act, as added by the Labor Management Relations Act, 29 U. S. C. § 164 (b), which states:

“Nothing in this Act shall be construed as authorizing the execution or application of agreements requiring membership1 in a labor organization as a condition of employment in any State or Territory *427in. which such execution or application is prohibited by State or Territorial law.”

Congress added this section to the Act “to forestall the inference that federal policy was to be exclusive.” Algoma Plywood Co., supra, at 314. Section 14 (b) “was designed to prevent other sections of the Act from completely extinguishing state power over certain union-security arrangements. ... It was desired to 'make certain’ that § 8 (a) (3) could not ‘be said to authorize arrangements of this sort in States where such arrangements were contrary to the State policy.’ ” Retail Clerks v. Schermerhorn, 373 U. S. 746, 751 (1963), quoting H. R. Conf. Rep. No. 510, supra, at 60.

To summarize, §§ 8 (a) (3) and 14(b) together exhaust the federal interest in the types of union-security agreements employers and unions may make. The closed shop is absolutely prohibited. And any lesser security arrangement, though consistent with the federal interest, is sanctioned only if it harmonizes with state policy.

It is undisputed that Texas law forbids union-shop agreements.4 The issue presented by this case, then, is whether this Texas law may extend to bar the security provision contained in the collective-bargaining agreement between the petitioners and the respondent. The petitioners contend that the applicability of state right-to-work laws depends upon where the work is to be performed. They conclude that because the employees in question perform 80% to 90% of their work on the high seas, the federal policy “favoring” union-shop provisions should prevail.5 The Government as amicus curiae, *428agreeing with the petitioners, asserts that Texas does not have a substantial enough interest to apply its labor policies since little work is performed within its borders. The respondent, in turn, relying upon the decisions of the District Court and the Court of Appeals, claims that the place of hiring is the key to analyzing the choice-of-law problem.

The language of § 14 (b) provides no clear guidance for determining whose law should prevail in a multijurisdictional situation. Section 14 (b) does prescribe a threshold: In order to apply its right-to-work laws, a State must be the place of “execution” or “application” of the union-security agreement. “Execution” and “application” are, however, broadly inclusive nouns. It is hardly conceivable that a State would wish to enforce its right-to-work laws unless the collective-bargaining agreement was iii some sense either executed or applied in the State. Yet clearly each of a number of States in a multistate situation could plausibly argue that it is the situs of “application” or “execution.” The State *429of “execution,” for example, could be considered to be either the State where the contract was signed or the State where the terms of the contract are to be carried out. “Application” could refer either to the hiring process or the performance of the work. In short, there is no intelligent way to infer from § 14 (b)’s expansive language which of a number of arguably relevant aspects of the employment relationship should be deemed the dispositive contact for deciding which State’s law is to apply.

The specific legislative history of § 14 (b) is of no greater aid in resolving the dilemma.6 Congress simply did not address the choice-of-law problems that would inevitably arise in multistate work force situations. We are, however, not entirely without signposts. When Congress legislated with respect to union-security agreements in 1947, it did not write on a clean slate, for few issues in American labor history had been as controverted as the moral legitimacy and, indeed, the legality, of union-security agreements. It is unnecessary for the purpose of this case to review the history of that long controversy.7 It is sufficient only to realize that Con*430gress did not resolve it, but instead left each individual State free to outlaw union-security agreements in the interest of a perceived policy of keeping industrial relations more individualistic, open, and free.

Although apparently no recorded legislative history exists to interpret the design of the Texas Legislature, the language of the statutes suggests that their principal purpose was, indeed, to democratize the hiring process. The Preamble of Public Policy contained in Art. 5154a, § 1, Tex. Rev. Civ. Stat. Ann. (1971), states in part:

“Because of the activities of labor unions affecting the economic conditions of the country and the State, entering as they do into practically every business and industrial enterprise, it is the sense of the Legislature that such organizations affect the public interest and are charged with a public use. The working man, unionist or non-unionist, must he protected. The right to work is the right to live.” (Emphasis added.)

Article 5207a states in part:

“Section 1. The inherent right of a person to work and bargain freely with his employer, individually or collectively, for terms and conditions of his employment shall not be denied or infringed by law, or by any organization of whatever nature.

“Sec. 2. No person shall be denied employment on account of membership or nonmembership in a labor union.”

Finally, Art. 5154g, § 1, states:

“It is hereby declared to be the public policy of the State of Texas that the right of persons to work shall not be denied or abridged on account of mem*431bership or non-membership in any labor union or labor organization and that in the exercise of such rights all persons shall be free from threats, force, intimidation or coercion.”

Each of these passages bespeaks an interest in a free hiring process and in preserving the freedom of the working man or woman to pursue and continue in employment, unhindered by coerced but unwanted union association.8

In Lunsford v. City of Bryan, 297 S. W. 2d 115, 117 (1957), the Supreme Court of Texas interpreted these statutes: “The intent seems obvious to protect employees in the exercise of the right of free choice of joining or not joining a union. The purpose of the statute is to afford equal opportunity to work to both classes of employees.” This authoritative state judicial interpretation thus confirms what seems manifest from the language of the statutes: Texas’ right-to-work laws are concerned with the process by which employees are hired and the conditions which, after their hiring, may burden their employment.

In the light of these purposes I agree with the District Court and the Court of Appeals that the laws of Texas govern the union-security agreement in this case. It is true that a number of States might legitimately assert an interest in the hiring process. The State where the employees reside, the State where the conditions of employment were negotiated, and the State where the hiring decision actually took place all have their claims. I. believe, however, that the State where *432the hiring actually takes place is, so far as the issue now before us goes, the most relevant jurisdiction for choice-of-law purposes, and that State in this case is Texas.9

In the first place, it seems clear that the State where the hiring actually takes place is the State most deeply concerned with the conditions of hire. The policy of a State such as Texas, which favors unrestricted hiring, will be seriously undermined when union-security agreements control the hiring that takes place within its jurisdiction. Moreover, the State where the hiring actually occurs normally provides the bulk of the work force from which the employees are drawn. And while a rule designating the laws of the State where the bargaining agreement was negotiated would provide for ease of application, it would also encourage forum shopping by both unions and management seeking the sanction of state laws that would most favor their interests.

Against this analysis, both the Government, as amicus, and the petitioners contend that job situs should be the determining factor in applying right-to-work laws. The parties do not explain, however, what the relevance of job situs is to laws that concern themselves exclusively *433with the hiring process. Cf. A. Von Mehren & D. Traut-man, The Law of Multistate Problems 76-79, 102-105 (1965). Indeed, the place where work is actually performed is probably the least relevant factor in the entire employment relationship for resolving conflicts over the legality of union-security agreements. It is undeniable, as the petitioners point out, that the job situs is where most of the day-to-day contact between employer and employees occurs. But right-to-work laws do not reflect a concern with job safety, or work rules, or hours of employment, or grievance procedures, or other similar conditions of employment with which the jurisdiction where the work actually takes place is legitimately concerned.10 *434A job situs test for the resolution of which State’s union-security law applies is, therefore, so arbitrary in my view as to approach irrationality.

II

But even if I could agree with the petitioners that the jobsite is the critical factor in determining what law should control the legality of union-security agreements, I would still find the laws of Texas applicable in this case. Quite simply, the employees here involved clearly perform a larger share of their employment duties in Texas than in any other State.

By contrast, the petitioners perceive this case as presenting a vertical conflict between Texas law and “federal law.” If the law of the jurisdiction where the work is performed controls, then, according to that perception, the “federal” rule ipso facto prevails, since 80% to 90% of the work is performed on the high seas.

The petitioners suggest alternative and somewhat inconsistent theories to justify the intrusion of “federal law” into this case. The first is that the high seas are, like the District of Columbia, a federal territory over which Congress exercises exclusive, pre-emptive jurisdiction.11 This theory is untenable. Congress undoubtedly has power under the Admiralty Clause, Art. Ill, § 2, to *435pre-empt the entire field of maritime law. E. g., Panama R. Co. v. Johnson, 264 U. S. 375, 386 (1924); Knickerbocker Ice Co. v. Stewart, 253 U. S. 149, 160 (1920); see G. Gilmore & C. Black, The Law of Admiralty 45-47 (1975); Currie, Federalism and the Admiralty: “The Devil’s Own Mess,” 1960 Sup. Ct. Rev. 158, 158-165. It has exercised this power liberally to regulate, for example, various aspects of maritime employment. See 46 U. S. C. §§ 563-568. Nevertheless, as Mr. Justice Frankfurter stated for the Court in Romero v. International Terminal Operating Co., 358 U. S. 354, 373 (1959):

“Although the corpus of admiralty law is federal in the sense that it derives from the implications of Article III evolved by the courts, to claim that all enforced rights pertaining to matters maritime are rooted in federal law is a destructive oversimplification of the highly intricate interplay of the States and the National Government in their regulation of maritime commerce. It is true that state law must yield to the needs of a uniform federal maritime law when this Court finds inroads on a harmonious system. But this limitation still leaves the States a wide scope.”

It is unnecessary here to delineate the “wide scope” within which the States may legislate about things maritime. To refute the notion that the high seas are a species of federal enclave, it is sufficient to point out that the Court has found state legislation pre-empted only when the nature of the problem required the application of a uniform rule or when the state law unduly hampered maritime commerce. See, e. g., Askew v. American Waterways Operators, Inc., 411 U. S. 325, 337-344 (1973); Kossick v. United Fruit Co., 365 U. S. 731, 738-739 (1961); Huron Cement Co. v. Detroit, 362 U. S. 440, 444 (1960). The Court has never struck *436down a state law on the ground that the States are juris-dictionally incompetent to legislate over matters that occur within the ocean “territory.”

The petitioners appear also to argue, however, that even if the high seas are not a territory over which Congress exercises exclusive lawmaking power, the Texas rule outlawing union shops must fall because the Federal Government has pre-empted the field of maritime labor relations. Cf. Southern Pacific Co. v. Jensen, 244 U. S. 205, 216 (1917); Currie, supra, at 165 passim. It is true that Congress has deeply involved itself in the affairs of seamen. Federal maritime law covers, among other things, maritime liens for the collection of wages, 46 U. S. C. § 953, the rights of seamen to receive at every port half of unpaid wages earned, 46 U. S. C. § 597, and double payment of wages withheld without sufficient cause, 46 U. S. C. § 596, physical qualifications and requirements for seamen, 46 U. S. C. § 672, and disciplinary problems, 46 U. S. C. §§ 701-710.

Despite this manifest federal interest in many' aspects of the maritime employment relationship, I think that Texas law still controls. Section 14 (b) on its face clearly settles any apparent conflict between state and federal law in favor of the state rule. In enacting § 14 (b) Congress concluded that diversity in the area of union-security agreements would compromise no federal interest. “By making the matter one of state law, Congress has not only authorized multiformity on the subject, but practically guaranteed it.” Motor Coach Employees v. Lockridge, 403 U. S. 274, 317 (1971) (White, J., dissenting).

When Congress has in the past determined that the nature of an interstate industry requires application of a uniform rule to govern union-security agreements, it has not hesitated to act. For example, before 1951 the Railway Labor Act, 45 U. S. C. § 152 Fifth, prohibited the *437union shop. Fully aware of § 14 (b) of the National Labor Relations Act, Congress amended the Railway Labor Act in 1951 to permit railroad carriers and their employees to enter into union-shop agreements “[notwithstanding any other provisions of this [Act], or of any other statute or law of the United States, or Territory thereof, or of any State . . . 45 U. S. C. § 152 Eleventh. See Railway Employes’ Dept. v. Hanson, 351 U. S. 225 (1956). That Congress has not similarly legislated for the maritime industry is compelling evidence that it finds compatible with federal interests diversity among the States as to permissible union-security agreements.

In conclusion, I believe that the place of hiring is the critical factor in determining the choice of law for union-security agreements. But even if the place where the work is to be performed is the criterion, Texas law should still be applied, since under this collective-bargaining agreement more work is performed in that State than in any other, and Congress has refrained from either establishing or indicating a need for a uniform rule to the contrary in maritime employment. I would, therefore, affirm the judgment before us.

concurring in the judgment.

Although I concur in the judgment of the Court, I do not think it necessary to determine in this case whether a “job situs” test is appropriate or required generally. The only issue before the Court is whether federal or state law should apply to the employment contracts of maritime workers whose job situs is the high seas and who thereby enjoy a special status. As noted by Judge Ainsworth, writing for the six dissenting members of the Court of Appeals:

“[Sjeamen have traditionally maintained an exceptional status in regard to the regulation and control of their employment, and . . . section 14 (b) cannot reasonably be construed to remove them from that category. Seamen, particularly the type of blue-water seamen involved here, as wards of admiralty have been accorded a special status and protection under federal maritime law unknown to state law in the domain of the master-servant relationship. Unlike the land-based worker, the seaman’s employ*422ment and all of the rights and restrictions flowing therefrom, are determined by federal statutory and admiralty law, not state law. . . .

. . The consistent and traditional control by federal law of every phase of maritime employment relationships and contracts refutes the proposition that [respondent’s] contacts with Texas justify injecting state law into federal maritime affairs.” 504 F. 2d 272, 284-286 (CA5 1974) (footnotes omitted).

I join in reversing the judgment of the Court of Appeals, as I do not believe § 14 (b) can be construed reasonably to apply to these seamen.

delivered the opinion of the Court.

Section 8 (a) (3) of the National Labor Relations Act, 49 Stat. 452, as amended, 61 Stat. 140, 29 U. S. C. § 158 (a)(3), permits employers as a matter of federal law to enter into agreements with unions to establish union or agency shops.1 Section 14 (b) of the Act, 61 Stat. 151, 29 U. S. C. § 164 (b), however, allows individual States and Territories to exempt themselves from § 8 (a) (3) and to enact so-called “right-to-work” laws prohibiting union or agency shops.2 We must de*410cide whether, under § 14 (b), Texas’ right-to-work laws can void an agency-shop agreement covering unlicensed seamen who, while hired in Texas and having a number of other contacts with the State, spend the vast majority of their working hours on the high seas.

I

Petitioners (hereinafter Union) 3 represent the unlicensed seamen who work on respondent employer’s oil tankers, In November 1969 the Union and respondent entered into a collective-bargaining agreement which provided for an agency shop: “For the duration of the Agreement all employees hired shall, as a condition of employment, become members of the Union and/ or in the alternative pay the regular union dues and initiation fees within 31 days from the employment date.” App. 281. Almost two years after entering into the agreement, respondent filed suit in the United States District Court for the Eastern District of Texas under § 301 of the Labor Management Relations Act, 61 Stat. 156, 29 U. S. C. § 185, claiming that the agency-shop provision was invalid and unenforceable because it violated Texas’ right-to-work laws.4

Uncontested evidence was presented at trial concerning the relevant locations of various aspects of the rela*411tionship between the Union, the respondent, and the seamen. Because this evidence bears heavily on the contentions of the parties, we shall summarize it in some detail. Respondent is a division of Mobil Oil Corp., a New York corporation, and operates a fleet of eight oceangoing tankers which transport respondent’s petroleum products from Texas to Atlantic coast ports. Respondent is headquartered in Beaumont, Tex., and maintains its personnel records there. Sixty percent of the applications to be unlicensed seamen on respondent’s ships are made in Beaumont and 40% in New York. The final hiring decisions are made in Beaumont. Of the 289 unlicensed seamen who are employed to man the tankers, 123 maintain residence in Texas, and 60 in New York.5 One hundred and fifty-two of the seamen list Beaumont as their shipping port — a designation that determines travel allowances to and from a seaman’s residence — and the remainder list either New York or Providence, R. I. Seamen can elect to be paid their wages aboard ship, to have their paychecks sent from the Beaumont office to designated recipients, or to use a combination of these two schemes. The collective-bargaining agreement whose agency-shop provision is at issue here was negotiated and executed in New York. It was re-executed in Texas.

A typical trip by one of respondent’s tankers from Beaumont, the Texas port, to Providence or New York, the Atlantic ports, takes from 4% to 5 days. Loading and unloading in port takes from 18 to 30 hours. No more than 10% to 20% of the seamen’s work time is spent within the territorial bounds of Texas.

Based on the above evidence, fully reflected in its *412findings of fact, the District Court concluded that “[t]he acts performed in the State of Texas in the administration and performance of the collective bargaining agreement are such that the State of Texas is intimately concerned with the collective bargaining agreement and with the employees working thereunder.” App. 29. Relying on this “intimate concern,” the court held that the Texas right-to-work laws were applicable under § 14 (b) and that the agency-shop provision was therefore void and unenforceable.

A three-member division of the United States Court of Appeals for the Fifth Circuit, one judge dissenting, reversed. 483 F. 2d 603 (1973). The court concluded that the Texas right-to-work laws could not apply since the employees’ principal job situs is not in Texas but rather is on the high seas. On rehearing en banc the full court, over the dissent of six of its members, vacated the division opinion and affirmed the judgment of the District Court. 504 F. 2d 272 (1974). The court identified and analyzed the interests that Texas has in the employment relationship at issue, placing special stress on the fact that all final hiring decisions take place in Texas. It held that “the federal labor legislation, the predominance of Texas contacts over any other jurisdiction, and the significant interest which Texas has in applying its right to work law to this employment relationship warrant application of the Texas law and, consequently, invalidation of the agency shop provision.” Id., at 275. We granted certiorari, 423 U. S. 820 (1975), and we now reverse.

II

All parties are agreed that the central inquiry in this case is whether § 14 (b) permits the application of Texas’ right-to-work laws to the agency-shop provision in the collective-bargaining agreement between the *413Union and respondent.6 Only if it is to be so read is the agency-shop provision unenforceable.7 The parties are similarly agreed that a State can apply its right-to-work laws only with respect to employment relationships with which the State has adequate contact. The crux of the differences between the parties concerns whether the contacts between Texas and the employment relationship in this case are sufficient to come under § 14 (b).

The Union, as well as the United States as amicus curiae, argues that the nature of the concerns at which § 14 (b) is directed mandates that job situs be the controlling factor in determining the applicability of § 14 (b), and that since in this case the employees’ principal job situs is on the high seas — outside the territorial bounds of the State — the agency-shop provision at issue is valid. Respondent contends that “[t]he sufficiency of a state’s interest in applying its law is to be determined by looking to the whole employment relationship.” Brief for Respondent 15. Giving weight to all the contacts between Texas and the employment relationship, see supra, at 410-411, respondent concludes that Texas can validly apply its right-to-work laws under *414§ 14 (b). A third approach, the one adopted by the dissenting opinion in this case, is that the location of the hiring process should be determinative of the applicability of a State’s right-to-work laws. Under this test, also, Texas’ contacts in this case would be sufficient to apply its laws.

In light of what we understand Congress’ concerns in both § 8 (a) (3) and § 14 (b) to have been, we conclude that it is the employees’ predominant job situs rather than a generalized weighing of factors or the place of hiring that triggers the operation of § 14 (b). We hold that under' § 14 (b), right-to-work laws cannot void agreements permitted by § 8 (a) (3) when the situs at which all the employees covered by the agreement perform most of their work is located outside of a State having such laws.

Under § 8 (3) of the Wagner Act, enacted in 1935, closed shops, union shops, and agency shops were all permitted. But in 1947, in § 8 (a) (3), as added by the Taft-Hartley Act, Congress reacted to widespread abuses of closed-shop agreements by banning such arrangements.8 Union and agency shops were still permitted, however, by §8 (a)(3). That provision makes employment discrimination in favor of or against labor unions an unfair labor practice but contains the following proviso which we have held to apply to agency shops as well as union shops, NLRB v. General Motors Corp., 373 U. S. 734 (1963):

“Provided, That nothing in this subchapter or in any other statute of the United States, shall preclude an employer from making an agreement with *415a labor organization ... to require as a condition of employment membership therein on or after the thirtieth day following the beginning of such employment or the effective date of such agreement, whichever is the later . . . .”

While permitting agency- and union-shop agreements, however, Congress provided certain safeguards for employees who were subject to such agreements. Thus a second proviso to § 8 (a) (3) warns:

“[N]o employer shall justify any discrimination against an employee for nonmembership in a labor organization (A) if he has reasonable grounds for believing that such membership was not available to the employee on the same terms and conditions generally applicable to other members, or (B) if he has reasonable grounds for believing that membership was denied or terminated for reasons other than the failure of the employee to tender the periodic dues and the initiation fees uniformly required as a condition of acquiring or retaining membership.”

Like its decision to ban closed-shop agreements, Congress’ decision in § 8 (a) (3) to provide these safeguards reflects a concern with compulsory unionism. But, in stark contrast to closed-shop agreements, these safeguards and the agency- or union-shop agreements to which they apply are not focused on the hiring process. Rather, they are directed at conditions that must be fulfilled by an employee only after he is already hired, at least 30 days after he is already working at the job-site.9 Moreover, quite apart from the safeguards that it *416provided, Congress’ decision to allow union-security agreements at all reflects its concern that, at least as a matter of federal law, the parties to a collective-bargaining agreement be allowed to provide that there be no employees who are getting the benefits of union representation without paying for them. Again,, the focus of this concern is not the hiring process, but rather the benefits to be derived from union representation during the period of employment — while the employee is on the job. Thus, the Senate Committee Report on what became the Taft-Hartley Act observed that § 8 (a)(3) gives “employers and unions who feel that [union-security] agreements promoted stability by eliminating 'free riders’ the right to continue such arrangements.” S. Rep. No. 105, 80th Cong., 1st Sess., 7 (1947), 1 Leg. Hist. 413. “Congress recognized that in the absence of a union-security provision 'many employees sharing the benefits of what unions are able to accomplish by collective bargaining will refuse to pay their share of the cost.’ S. Rep. No. 105, 80th Cong., 1st Sess., p. 6, 1 Leg. Hist. L. M. R. A. 412.” NLRB v. General Motors Corp., supra, at 740-741.

In short, insofar as it deals with union-security agreements less onerous than the closed-shop agreement, § 8 (a)(3) focuses in both effect and purpose on post-hiring conditions, conditions which have a major impact on the job situs.

While § 8 (a) (3) articulates a national policy that certain union-security agreements are valid as a matter of federal law, § 14 (b) reflects Congress’ decision that any *417State or Territory that wishes to may exempt itself from that policy. Section 14 (b) allows a State or Territory to ban agreements “requiring membership in a labor organization as a condition of employment.” 10 We have recognized that with respect to those state laws which § 14 (b) permits to be exempted from § 8 (a)(3)’s national policy “[t]here is . . . conflict between state and federal law; but it is a conflict sanctioned by Congress with directions to give the right of way to state laws . . . .” Retail Clerks v. Schermerhorn, 375 U. S. 96, 103 (1963). The question here, of course, is whether Texas’ contacts with this employment relationship are adequate to call into play § 14 (b)’s mandated deference to state law.

Section 14 (b) simply mirrors that part of § 8 (a) (3) which focuses on posi-hiring conditions of employment. As its language reflects, § 14 (b) was designed to make clear that § 8 (a) (3) left the States free to pursue “their own more restrictive policies in the matter of union-security agreements.” Algoma Plywood Co. v. Wisconsin Board, 336 U. S. 301, 314 (1949). Since § 8 (a)(3) already prohibits the closed shop, the more restrictive policies that § 14 (b) allows the States to enact relate not to the hiring process but rather to conditions that would come into effect only after an individual is hired. It is evident, then, that § 14 (b)’s primary concern is with state regulation of the posi-hiring employer-employee-union relationship. And the center of the post-hiring relationship is the job situs, the place where the work that is the very raison d’etre of the relationship is performed.

The centrality of job situs to Congress’ concern in § 14 (b) is also suggested by the House Committee Report on the bill that contained the substance of what was *418finally enacted as § 14(b). That report reflects the House’s intent that agreements providing for agency or union shops would be valid “only if they are valid under the laws of any State in which they are to be performed.” H. R. Rep. No. 245, 80th Cong., 1st Sess., 34 (1947), 1 Leg. Hist. 325. Where an agreement is “performed” may be open to some debate, but we think the most reasonable reading of the phrase is that union-security agreements are “performed” on the job situs. Thus, the import of the House Report is that the committee viewed what became § 14 (b) as allowing a State to ban agreements calling for work to be performed within the State. While the Taft-Hartley Act as finally enacted does not contain the precise wording of the House bill, there is no indication that any language changes were designed to alter this focus on the place of performance.

Whether taken separately or together, the place of hiring and the other factors on which respondent relies — the employees’ place of residence, the locale of personnel records, the place at which payroll checks are written, etc. — are not nearly as central to the concerns of § 14 (b) as the employees’ job situs. And, because of this close relationship between § 14 (b) and job situs we conclude that § 14 (b) does not allow enforcement of right-to-work laws with regard to an employment relationship whose principal job situs is outside of a State having such laws.

Two practical considerations bolster our conclusion that the employees’ predominant job situs should determine the applicability of a State’s right-to-work laws under § 14 (b). First, the use of a job situs test will minimize the possibility of patently anomalous extraterritorial applications of any given State’s right-to-work laws. Use of a job situs test will insure that the laws of a State with a continuing and current relationship *419with the employees in question will govern the validity vel non of any union-shop or agency-shop provision. On the other hand, if place of hiring were to be the determinative factor, Texas, for instance, could apply its right-to-work laws to employees who work solely in Connecticut simply because the relevant hiring decisions were made — perhaps many years ago — in Texas. We cannot believe that it was Congress’ purpose in passing § 14 (b) to sanction such a result.

A test such as the one adopted by the Court of Appeals that evaluates all of a jurisdiction’s employment relationship contacts in order to determine the applicability of its right-to-work laws under § 14 (b) might not result in irrational extraterritorial applications. But such a test does suffer the disadvantages of being both less predictable and more difficult of application than a job situs test. Under a job situs test, parties entering a collective-bargaining agreement will easily be able to determine in virtually all situations whether a union- or agency-shop provision is valid. By contrast, bargaining parties would often be left in a state of considerable uncertainty if they were forced to identify and evaluate all the relevant contacts of a jurisdiction in order to determine the potential validity of a proposed union-security provision. The unpredictability that such a test would inject into the bargaining relationship, as well as the burdens of litigation that would result from it, make us unwilling to impute to Congress any intent to adopt such a test.11

*420Ill

Having concluded that predominant job situs is the controlling factor in determining whether, under § 14 (b), a State can apply its right-to-work laws to a given employment relationship, the disposition of this case is clear. Because most of the employees’ work is done on the high seas, outside the territorial bounds of the State of Texas, Texas’ right-to-work laws cannot govern the validity of the agency-shop provision at issue here. It is immaterial that Texas may have more contacts than any other State with the employment relationship in this case, since there is no reason to conclude under § 14 (b) that in every employment situation some State or Territory’s law with respect to union-security agreements must be applicable.12 Federal policy favors permitting such agreements unless a State or Territory with a sufficient interest in the relationship expresses a contrary policy via right-to-work laws. It is therefore fully consistent with national labor policy to conclude, if the predominant job situs is outside the boundary of any State, that no State *421has a sufficient interest in the employment relationship and that no State’s right-to-work laws can apply.

Accordingly, the judgment of the Court of Appeals is reversed.

So ordered.

Mr. Chief Justice Burger concurs in the judgment.