announced the judgment of the Court and delivered an opinion, in which Justice Marshall, Justice Blackmun, and Justice Stevens joined.

The question presented is whether the assignment by Congress to bankruptcy judges of the jurisdiction granted in 28 U. S. C. § 1471 (1976 ed., Supp. IV) by § 241(a) of the Bankruptcy Act of 1978 violates Art. Ill of the Constitution.

I

A

In 1978, after almost 10 years of study and investigation, Congress enacted a comprehensive revision of the bank*53ruptcy laws. The Bankruptcy Act of 1978 (Act)1 made significant changes in both the substantive and procedural law of bankruptcy. It is the changes in the latter that are at issue in this case.

Before the Act, federal district courts served as bankruptcy courts and employed a “referee” system. Bankruptcy proceedings were generally conducted before referees,2 except in those instances in which the district court elected to withdraw a case from a referee. See Bkrtcy. Rule 102. The referee’s final order was appealable to the district court. Bkrtcy. Rule 801. The bankruptcy courts were vested with “summary jurisdiction” — that is, with jurisdiction over controversies involving property in the actual or constructive possession of the court. And, with consent, the bankruptcy court also had jurisdiction over some “plenary” matters — such as disputes involving property in the possession of a third person.

The Act eliminates the referee system and establishes “in each judicial district, as an adjunct to the district court for such district, a bankruptcy court which shall be a court of record known as the United States Bankruptcy Court for the district.” 28 U. S. C. § 151(a) (1976 ed., Supp. IV). The judges of these courts are appointed to office for 14-year terms by the President, with the advice and consent of the Senate. §§152, 153(a) (1976 ed., Supp IV). They are subject to removal by the “judicial council of the circuit” on account of “incompetency, misconduct, neglect of duty or physical or mental disability.” § 153(b) (1976 ed., Supp. IV). In addition, the salaries of the bankruptcy judges are set by statute and are subject to adjustment under the Federal Salary Act, 2 U. S. C. §§351-361 (1976 ed. and Supp. IV). 28 U. S. C. §154 (1976 ed., Supp. IV).

*54The jurisdiction of the bankruptcy courts created by the Act is much broader than that exercised under the former referee system. Eliminating the distinction between “summary” and “plenary” jurisdiction, the Act grants the new courts jurisdiction over all “civil proceedings arising under title 11 [the Bankruptcy title] or arising in or related to cases under title 11.” 28 U. S. C. § 1471(b) (1976 ed., Supp. IV) (emphasis added).3 This jurisdictional grant empowers bankruptcy courts to entertain a wide variety of cases involving claims that may affect the property of the estate once a petition has been filed under Title 11. Included within the bankruptcy courts’ jurisdiction are suits to recover accounts, controversies involving exempt property, actions to avoid transfers and payments as preferences or fraudulent conveyances, and causes of action owned by the debtor at the time of the petition for bankruptcy. The bankruptcy courts can hear claims based on state law as well as those based on federal law. See 1 W. Collier, Bankruptcy ¶3.01, pp. 3-47 to 3-48 (15th ed. 1982).4

*55The judges of the bankruptcy courts are vested with all of the “powers of a court of equity, law, and admiralty,” except that they “may not enjoin another court or punish a criminal contempt not committed in the presence of the judge of the court or warranting a punishment of imprisonment.” 28 U. S. C. § 1481 (1976 ed., Supp. IV). In addition to this broad grant of power, Congress has allowed bankruptcy judges the power to hold jury trials, § 1480; to issue declaratory judgments, § 2201; to issue writs of habeas corpus under certain circumstances, §2256; to issue all writs necessary in aid of the bankruptcy court’s expanded jurisdiction, §451 (1976 ed. and Supp. IV); see 28 U. S. C. §1651; and to issue any order, process or judgment that is necessary or appropriate to carry out the provisions of Title 11, 11 U. S. C. § 105(a) (1976 ed., Supp. IV).

The Act also establishes a special procedure for appeals from orders of bankruptcy courts. The circuit council is empowered to direct the chief judge of the circuit to designate panels of three bankruptcy judges to hear appeals. 28 U. S. C. §160 (1976 ed., Supp. IV). These panels have jurisdiction of all appeals from final judgments, orders, and decrees of bankruptcy courts, and, with leave of the panel, of interlocutory appeals. § 1482. If no such appeals panel is designated, the district court is empowered to exercise appellate jurisdiction. § 1334. The court of appeals is given jurisdiction over appeals from the appellate panels or from the district court. § 1293. If the parties agree, a direct appeal to the court of appeals may be taken from a final judgment of a bankruptcy court. § 1293(b).5

*56The Act provides for a transition period before the new provisions take full effect in April 1984. §§ 401-411, 92 Stat. 2682-2688. During the transition period, previously existing bankruptcy courts continue in existence. § 404(a), 92 Stat. 2683. Incumbent bankruptcy referees, who served 6-year terms for compensation subject to adjustment by Congress, are to serve as bankruptcy judges until March 31, 1984, or until their successors take office. § 404(b), 92 Stat. 2683.6 During this period they are empowered to exercise essentially all of the jurisdiction and powers discussed above. See §§ 404, 405, 92 Stat. 2683-2685. See generally 1 Collier, supra, ¶¶ 7.04-7.05, pp. 7-23 to 7-65. The procedure for taking appeals is similar to that provided after the transition period. See § 405(c)(1), 92 Stat. 2685.7

B

This case arises out of proceedings initiated in the United States Bankruptcy Court for the District of Minnesota after appellant Northern Pipeline Construction Co. (Northern) filed a petition for reorganization in January 1980. In March 1980 Northern, pursuant to the Act, filed in that court a suit against appellee Marathon Pipe Line Co. (Marathon). Appellant sought damages for alleged breaches of contract and warranty, as well as for alleged misrepresentation, coercion, and duress. Marathon sought dismissal of the suit, on the ground that the Act unconstitutionally conferred Art. Ill ju*57dicial power upon judges who lacked life tenure and protection against salary diminution. The United States intervened to defend the validity of the statute.

The Bankruptcy Judge denied the motion to dismiss. 6 B. R. 928 (1980). But on appeal the District Court entered an order granting the motion, on the ground that “the delegation of authority in 28 U. S. C. § 1471 to the Bankruptcy Judges to try cases which are otherwise relegated under the Constitution to Article III judges” was unconstitutional. Both the United States and Northern filed notices of appeal in this Court.8 We noted probable jurisdiction. 454 U. S. 1029 (1981).9

II

A

Basic to the constitutional structure established by the Framers was their recognition that “[t]he accumulation of all powers, legislative, executive, and judiciary, in the same hands, whether of one, a few, or many, and whether hereditary, self-appointed, or elective, may justly be pronounced the very definition of tyranny.” The Federalist No. 47, p. 300 (H. Lodge ed. 1888) (J. Madison). To ensure against such tyranny, the Framers provided that the Federal Government would consist of three distinct Branches, each to exercise one of the governmental powers recognized by the Framers as inherently distinct. “The Framers regarded the checks and balances that they had built into the tripartite Federal Government as a self-executing safeguard against the encroachment or aggrandizement of one branch at the *58expense of the other.” Buckley v. Valeo, 424 U. S. 1, 122 (1976) (per curiam).

The Federal Judiciary was therefore designed by the Framers to stand independent of the Executive and Legislature — to maintain the checks and balances of the constitutional structure, and also to guarantee that the process of adjudication itself remained impartial. Hamilton explained the importance of an independent Judiciary:

“Periodical appointments, however regulated, or by whomsoever made, would, in some way or other, be fatal to [the courts’] necessary independence. If the power of making them was committed either to the Executive or legislature, there would be danger of an improper complaisance to the branch which possessed it; if to both, there would be an unwillingness to hazard the displeasure of either; if to the people, or to persons chosen by them for the special purpose, there would be too great a disposition to consult popularity, to justify a reliance that nothing would be consulted but the Constitution and the laws.” The Federalist No. 78, p. 489 (H. Lodge ed. 1888).

The Court has only recently reaffirmed the significance of this feature of the Framers’ design: “A Judiciary free from control by the Executive and Legislature is essential if there is a right to have claims decided by judges who are free from potential domination by other branches of government.” United States v. Will, 449 U. S. 200, 217-218 (1980).

As an inseparable element of the constitutional system of checks and balances, and as a guarantee of judicial impartiality, Art. Ill both defines the power and protects the independence of the Judicial Branch. It provides that “The judicial Power of the United States, shall be vested in one supreme Court, and in such inferior Courts as the Congress may from time to time ordain and establish.” Art. Ill, § 1. The inexorable command of this provision is clear and defi*59nite: The judicial power of the United States must be exercised by courts having the attributes prescribed in Art. III. Those attributes are also clearly set forth:

“The Judges, both of the supreme and inferior Courts, shall hold their Offices during good Behaviour, and shall, at stated Times, receive for their Services, a Compensation, which shall not be diminished during their Continuance in Office.” Art. Ill, §1.

The “good Behaviour” Clause guarantees that Art. Ill judges shall enjoy life tenure, subject only to removal by impeachment. United States ex rel. Toth v. Quarles, 350 U. S. 11, 16 (1955). The Compensation Clause guarantees Art. Ill judges a fixed and irreducible compensation for their services. United States v. Will, supra, at 218-221. Both of these provisions were incorporated into the Constitution to ensure the independence of the Judiciary from the control of the Executive and Legislative Branches of government.10 As we have only recently emphasized, “[t]he Compensation Clause has its roots in the longstanding Anglo-American tradition of an independent Judiciary,” 449 U. S., at 217, while the principle of life tenure can be traced back at least as far as the Act of Settlement in 1701, id., at 218. To be sure, both principles were eroded during the late colonial period, but that departure did not escape notice and indignant rejection by the Revolutionary generation. Indeed, the guarantees eventually included *60in Art. Ill were clearly foreshadowed in the Declaration of Independence, “which, among the injuries and usurpations recited against the King of Great Britain, declared that he had ‘made judges dependent on his will alone, for the tenure of their offices, and the amount and payment of their salaries.’” O’Donoghue v. United States, 289 U. S. 516, 531 (1933). The Framers thus recognized:

“Next to permanency in office, nothing can contribute more to the independence of the judges than a fixed provision for their support. ... In the general course of human nature, a power over a man’s subsistence amounts to a power over his will.” The Federalist No. 79, p. 491 (H. Lodge ed. 1888) (A. Hamilton) (emphasis in original).11

In sum, our Constitution unambiguously enunciates a fundamental principle — that the “judicial Power of the United States” must be reposed in an independent Judiciary. It commands that the independence of the Judiciary be jealously guarded, and it provides clear institutional protections for that independence.

B

It is undisputed that the bankruptcy judges whose offices were created by the Bankruptcy Act of 1978 do not enjoy the protections constitutionally afforded to Art. Ill judges. The bankruptcy judges do not serve for life subject to their continued “good Behaviour.” Rather, they are appointed for *6114-year terms, and can be removed by the judicial council of the circuit in which they serve on grounds of “incompetency, misconduct, neglect of duty, or physical or mental disability.” Second, the salaries of the bankruptcy judges are not immune from diminution by Congress. See supra, at 53. In short, there is no doubt that the bankruptcy judges created by the Act are not Art. Ill judges.

That Congress chose to vest such broad jurisdiction in non-Art. Ill bankruptcy courts, after giving substantial consideration to the constitutionality of the Act, is of course reason to respect the congressional conclusion. See Fullilove v. Klutznick, 448 U. S. 448, 472-473 (1980) (opinion of Burger, C. J.); Palmore v. United States, 411 U. S. 389, 409 (1973). See also National Ins. Co. v. Tidewater Co., 337 U. S. 582, 655 (1949) (Frankfurter, J., dissenting).12 But at the same time,

*62“[deciding whether a matter has in any measure been committed by the Constitution to another branch of government, or whether the action of that branch exceeds whatever authority has been committed, is itself a delicate exercise in constitutional interpretation, and is a responsibility of this Court as ultimate interpreter of the Constitution.” Baker v. Carr, 369 U. S. 186, 211 (1962).

With these principles in mind, we turn to the question presented for decision: whether the Bankruptcy Act of 1978 violates the command of Art. Ill that the judicial power of the United States must be vested in courts whose judges enjoy the protections and safeguards specified in that Article.

Appellants suggest two grounds for upholding the Act’s conferral of broad adjudicative powers upon judges unprotected by Art. III. First, it is urged that “pursuant to its enumerated Article I powers, Congress may establish legislative courts that have jurisdiction to decide cases to which the Article III judicial power of the United States extends.” Brief for United States 9. Referring to our precedents upholding the validity of “legislative courts,” appellants suggest that “the plenary grants of power in Article I permit Congress to establish non-Article III tribunals in ‘specialized areas having particularized needs and warranting distinctive treatment,’” such as the area of bankruptcy law. Ibid., quoting Palmore v. United States, supra, at 408. Second, appellants contend that even if the Constitution does require that this bankruptcy-related action be adjudicated in an Art. Ill court, the Act in fact satisfies that requirement. “Bank*63ruptcy jurisdiction was vested in the district court” of the judicial district in which the bankruptcy court is located, “and the exercise of that jurisdiction by the adjunct bankruptcy court was made subject to appeal as of right to an Article III court.” Brief for United States 12. Analogizing the role of the bankruptcy court to that of a special master, appellants urge us to conclude that this “adjunct” system established by Congress satisfies the requirements of Art. III. We consider these arguments in turn.

Ill

Congress did not constitute the bankruptcy courts as legislative courts.13 Appellants contend, however, that the bankruptcy courts could have been so constituted, and that as a result the “adjunct” system in fact chosen by Congress does not impermissibly encroach upon the judicial power. In advancing this argument, appellants rely upon cases in which we have identified certain matters that “congress may or may not bring within the cognizance of [Art. Ill courts], as it may deem proper.” Murray’s Lessee v. Hoboken Land & Improvement Co., 18 How. 272, 284 (1856).14 But when properly understood, these precedents represent no broad departure from the constitutional command that the judicial power of the United States must be vested in Art. III *64courts.15 Rather, they reduce to three narrow situations not subject to that command, each recognizing a circumstance in which the grant of power to the Legislative and Executive Branches was historically and constitutionally so exceptional that the congressional assertion of a power to create legislative courts was consistent with, rather than threatening to, the constitutional mandate of separation of powers. These precedents simply acknowledge that the literal command of Art. Ill, assigning the judicial power of the United States to courts insulated from Legislative or Executive interference, must be interpreted in light of the historical context in which the Constitution was written, and of the structural imperatives of the Constitution as a whole.

Appellants first rely upon a series of cases in which this Court has upheld the creation by Congress of non-Art. Ill “territorial courts.” This exception from the general prescription of Art. Ill dates from the earliest days of the Republic, when it was perceived that the Framers intended that as to certain geographical areas, in which no State operated as sovereign, Congress was to exercise the general powers of government. For example, in American Ins. Co. v. Canter, 1 Pet. 511 (1828), the Court observed that Art. IV bestowed upon Congress alone a complete power of government over *65territories not within the States that constituted the United States. The Court then acknowledged Congress’ authority to create courts for those territories that were not in conformity with Art. III. Such courts were

“created in virtue of the general right of sovereignty which exists in the government, or in virtue of that clause which enables Congress to make all needful rules and regulations, respecting the territory belonging to the United States. The jurisdiction with which they are invested ... is conferred by Congress, in the execution of those general powers which that body possesses over the territories of the United States. Although admiralty jurisdiction can be exercised in the states in those Courts, only, which are established in pursuance of the third article of the Constitution; the same limitation does not extend to the territories. In legislating for them, Congress exercises the combined powers of the general, and of a state government.” 1 Pet., at 546.

The Court followed the same reasoning when it reviewed Congress’ creation of non-Art. Ill courts in the District of Columbia. It noted that there was in the District

“no division of powers between the general and state governments. Congress has the entire control over the district for every purpose of government; and it is reasonable to suppose, that in organizing a judicial department here, all judicial power necessary for the purposes of government would be vested in the courts of justice.” Kendall v. United States, 12 Pet. 524, 619 (1838).16

*66Appellants next advert to a second class of cases — those in which this Court has sustained the exercise by Congress and the Executive of the power to establish and administer courts-martial. The situation in these cases strongly resembles the situation with respect to territorial courts: It too involves a constitutional grant of power that has been historically understood as giving the political Branches of Government extraordinary control over the precise subject matter at issue. Article I, § 8, els. 13, 14, confer upon Congress the power “[t]o provide and maintain a Navy,” and “[t]o make Rules for the Government and Regulation of the land and naval Forces.” The Fifth Amendment, which requires a presentment or indictment of a grand jury before a person may be held to answer for a capital or otherwise infamous crime, contains an express exception for “cases arising in the land or naval forces.” And Art. II, §2, cl. 1, provides that “The President shall be Commander in Chief of the Army and Navy of the United States, and of the Militia of the several States, when called into the actual Service of the United States.” Noting these constitutional directives, the Court in Dynes v. Hoover, 20 How. 65 (1857), explained:

“These provisions show that Congress has the power to provide for the trial and punishment of military and naval offences in the manner then and now practiced by civilized nations; and that the power to do so is given without any connection between it and the 3d article of the Constitution defining the judicial power of the United States; indeed, that the two powers are entirely independent of each other.” Id., at 79.17

*67Finally, appellants rely on a third group of cases, in which this Court has upheld the constitutionality of legislative courts and administrative agencies created by Congress to adjudicate cases involving “public rights.”18 The “public rights” doctrine was first set forth in Murray’s Lessee v. Hoboken Land & Improvement Co., 18 How. 272 (1856):

“[W]e do not consider congress can either withdraw from judicial cognizance any matter which, from its nature, is the subject of a suit at the common law, or in equity, or admiralty; nor, on the other hand, can it bring under the judicial power a matter which, from its nature, is not a subject for judicial determination. At the same time there are matters, involving public rights, which may be presented in such form that the judicial power is capable of acting on them, and which are susceptible of judicial determination, but which congress may or may not bring within the cognizance of the courts of the United States, as it may deem proper.” Id., at 284 (emphasis added).

This doctrine may be explained in part by reference to the traditional principle of sovereign immunity, which recognizes that the Government may attach conditions to its consent to be sued. See id., at 283-285; see also Ex parte Bakelite Corp., 279 U. S. 438, 452 (1929). But the public-rights doctrine also draws upon the principle of separation of powers, and a historical understanding that certain prerogatives were reserved to the political Branches of Government. The doctrine extends only to matters arising “between the Gov*68ernment and persons subject to its authority in connection with the performance of the constitutional functions of the executive or legislative departments,” Crowell v. Benson, 285 U. S. 22, 50 (1932), and only to matters that historically could have been determined exclusively by those departments, see Ex parte Bakelite Corp., supra, at 458. The understanding of these cases is that the Framers expected that Congress would be free to commit such matters completely to nonjudicial executive determination, and that as a result there can be no constitutional objection to Congress’ employing the less drastic expedient of committing their determination to a legislative court or an administrative agency. Crowell v. Benson, supra, at 50.19

The public-rights doctrine is grounded in a historically recognized distinction between matters that could be conclusively determined by the Executive and Legislative Branches and matters that are “inherently. . . judicial.” Ex parte Bakelite Corp., supra, at 458. See Murray’s Lessee v. Hoboken Land & Improvement Co., 18 How., at 280-282. For example, the Court in Murray’s Lessee looked to the law of England and the States at the time the Constitution was adopted, in order to determine whether the issue presented was customarily cognizable in the courts. Ibid. Concluding that the matter had not traditionally been one for judicial determination, the Court perceived no bar to Congress’ establishment of summary procedures, outside of Art. Ill courts, to collect a debt due to the Government from one of its customs agents.20 On the same premise, the Court in Ex *69parte Bakelite Corp., supra, held that the Court of Customs Appeals had been properly constituted by Congress as a legislative court:

“The full province of the court under the act creating it is that of determining matters arising between the Government and others in the executive administration and application of the customs laws. . . . The appeals include nothing which inherently or necessarily requires judicial determination, but only matters the determination of which may be, and at times has been, committed exclusively to executive officers .” 279 U. S., at 458 (emphasis added).21

The distinction between public rights and private rights has not been definitively explained in our precedents.22 Nor is it necessary to do so in the present cases, for it suffices to observe that a matter of public rights must at a minimum arise “between the government and others.” Ex parte Bakelite Corp., supra, at 451.23 In contrast, “the liability of *70one individual to another under the law as defined,” Crowell v. Benson, supra, at 51, is a matter of private rights. Our precedents clearly establish that only controversies in the former category may be removed from Art. Ill courts and delegated to legislative courts or administrative agencies for their determination. See Atlas Roofing Co. v. Occupational Safety and Health Review Comm’n, 430 U. S. 442, 450, n. 7 (1977); Crowell v. Benson, supra, at 50-51. See also Katz, Federal Legislative Courts, 43 Harv. L. Rev. 894, 917-918 (1930).24 Private-rights disputes, on the other hand, lie at the core of the historically recognized judicial power.

In sum, this Court has identified three situations in which Art. Ill does not bar the creation of legislative courts. In each of these situations, the Court has recognized certain exceptional powers bestowed upon Congress by the Constitution or by historical consensus. Only in the face of such an exceptional grant of power has the Court declined to hold the authority of Congress subject to the general prescriptions of Art. III.25

*71We discern no such exceptional grant of power applicable in the cases before us. The courts created by the Bankruptcy Act of 1978 do not lie exclusively outside the States of the Federal Union, like those in the District of Columbia and the Territories. Nor do the bankruptcy courts bear any resemblance to courts-martial, which are founded upon the Constitution’s grant of plenary authority over the Nation’s military forces to the Legislative and Executive Branches. Finally, the substantive legal rights at issue in the present action cannot be deemed “public rights.” Appellants argue that a discharge in bankruptcy is indeed a “public right,” similar to such congressionally created benefits as “radio station licenses, pilot licenses, or certificates for common carriers” granted by administrative agencies. See Brief for United States 34. But the restructuring of debtor-creditor relations, which is at the core of the federal bankruptcy power, must be distinguished from the adjudication of state-created private rights, such as the right to recover contract damages that is at issue in this case. The former may well be a “public right,” but the latter obviously is not. Appellant Northern’s right to recover contract damages to augment its estate is “one of private right, that is, of the liability of one *72individual to another under the law as defined.” Crowell v. Benson, 285 U. S., at 51.26

Recognizing that the present cases may not fall within the scope of any of our prior cases permitting the establishment of legislative courts, appellants argue that we should recognize an additional situation beyond the command of Art. Ill, sufficiently broad to sustain the Act. Appellants contend that Congress’ constitutional authority to establish “uniform Laws on the subject of Bankruptcies throughout the United States,” Art. I, §8, cl. 4, carries with it an inherent power to establish legislative courts capable of adjudicating “bankruptcy-related controversies.” Brief for United States 14. In support of this argument, appellants rely primarily upon a quotation from the opinion in Palmore v. United States, 411 U. S. 389 (1973), in which we stated that

“both Congress and this Court have recognized that. . . the requirements of Art. Ill, which are applicable where laws of national applicability and affairs of national concern are at stake, must in proper circumstances give way to accommodate plenary grants of power to Congress to legislate with respect to specialized areas having particularized needs and warranting distinctive treatment.” Id., 407-408.

Appellants cite this language to support their proposition that a bankruptcy court created by Congress under its Art. I *73powers is constitutional, because the law of bankruptcy is a “specialized area,” and Congress has found a “particularized need” that warrants “distinctive treatment.” Brief for United States 20-33.

Appellants’ contention, in essence, is that pursuant to any of its Art. I powers, Congress may create courts free of Art. Ill’s requirements whenever it finds that course expedient. This contention has been rejected in previous cases. See, e. g., Atlas Roofing Co. v. Occupational Safety and Health Review Comm’n, 430 U. S., at 450, n. 7; United States ex rel. Toth v. Quarles, 350 U. S. 11 (1955). Although the cases relied upon by appellants demonstrate that independent courts are not required for all federal adjudications, those cases also make it clear that where Art. Ill does apply, all of the legislative powers specified in Art. I and elsewhere are subject to it. See, e. g., Ex parte Bakelite Corp., 279 U. S., at 449; United States ex rel. Toth v. Quarles, supra; American Ins. Co. v. Canter, 1 Pet., at 546; Murray’s Lessee, 18 How., at 284. Cf. Crowell v. Benson, supra, at 51.

The flaw in appellants’ analysis is that it provides no limiting principle. It thus threatens to supplant completely our system of adjudication in independent Art. Ill tribunals and replace it with a system of “specialized” legislative courts. True, appellants argue that under their analysis Congress could create legislative courts pursuant only to some “specific” Art. I power, and “only when there is a particularized need for distinctive treatment.” Brief for United States 22-23. They therefore assert that their analysis would not permit Congress to replace the independent Art. Ill Judiciary through a “wholesale assignment of federal judicial business to legislative courts.” Ibid. But these “limitations” are wholly illusory. For example, Art. I, § 8, empowers Congress to enact laws, inter alia, regulating interstate commerce and punishing certain crimes. Art. I, § 8, els. 3, 6. On appellants’ reasoning Congress could provide for the adjudication of these and “related” matters by judges and *74courts within Congress’ exclusive control.27 The potential for encroachment upon powers reserved to the Judicial Branch through the device of “specialized” legislative courts is dramatically evidenced in the jurisdiction granted to the courts created by the Act before us. The broad range of questions that can be brought into a bankruptcy court because they are “related to cases under title 11,” 28 U. S. C. § 1471(b) (1976 ed., Supp. IV), see supra, at 54, is the clearest proof that even when Congress acts through a “specialized” court, and pursuant to only one of its many Art. I powers, appellants’ analysis fails to provide any real protection against the erosion of Art. Ill jurisdiction by the unilateral action of the political Branches. In short, to accept appellants’ reasoning, would require that we replace the principles delineated in our precedents, rooted in history and the Constitution, with a rule of broad legislative discretion that could effectively eviscerate the constitutional guarantee of an independent Judicial Branch of the Federal Government.28

*75Appellants’ reliance upon Palmore for such broad legislative discretion is misplaced. In the context of the issue decided in that case, the language quoted from the Palmore opinion, supra, at 72, offers no substantial support for appellants’ argument. Palmore was concerned with the courts of the District of Columbia, a unique federal enclave over which “Congress has . . . entire control ... for every purpose of government.” Kendall v. United States, 12 Pet., at 619. *76The “plenary authority” under the District of Columbia Clause, Art. I, §8, cl. 17, was the subject of the quoted passage and the powers granted under that Clause are obviously different in kind from the other broad powers conferred on Congress: Congress’ power over the District of Columbia encompasses the full authority of government, and thus, necessarily, the Executive and Judicial powers as well as the Legislative. This is a power that is clearly possessed by Congress only in limited geographic areas. Palmore itself makes this limitation clear. The quoted passage distinguishes the congressional powers at issue in Palmore from those in which the Art. Ill command of an independent Judiciary must be honored: where “laws of national applicability and affairs of national concern are at stake.” 411 U. S., at 408. Laws respecting bankruptcy, like most laws enacted pursuant to the national powers cataloged in Art. I, § 8, are clearly laws of national applicability and affairs of national concern. Thus our reference in Palmore to “specialized areas having particularized needs” referred only to geographic areas, such as the District of Columbia or territories outside the States of the Federal Union. In light of the clear commands of Art. Ill, nothing held or said in Palmore can be taken to mean that in every area in which Congress may legislate, it may also create non-Art. Ill courts with Art. Ill powers.

In sum, Art. Ill bars Congress from establishing legislative courts to exercise jurisdiction over all matters related to those arising under the bankruptcy laws. The establishment of such courts does not fall within any of the historically recognized situations in which the general principle of independent adjudication commanded by Art. Ill does not apply. Nor can we discern any persuasive reason, in logic, history, or the Constitution, why the bankruptcy courts here established lie beyond the reach of Art. III.

IV

Appellants advance a second argument for upholding the constitutionality of the Act: that “viewed within the entire ju*77dicial framework set up by Congress,” the bankruptcy court is merely an “adjunct” to the district court, and that the delegation of certain adjudicative functions to the bankruptcy court is accordingly consistent with the principle that the judicial power of the United States must be vested in Art. Ill courts. See Brief for United States 11-13, 37-45. As support for their argument, appellants rely principally upon Crowell v. Benson, 285 U. S. 22 (1932), and United States v. Raddatz, 447 U. S. 667 (1980), cases in which we approved the use of administrative agencies and magistrates as adjuncts to Art. Ill courts. Brief for United States 40-42. The question to which we turn, therefore, is whether the Act has retained “the essential attributes of the judicial power,” Crowell v. Benson, supra, at 51, in Art. III tribunals.29

The essential premise underlying appellants’ argument is that even where the Constitution denies Congress the power to establish legislative courts, Congress possesses the authority to assign certain factfinding functions to adjunct tribunals. It is, of course, true that while the power to adjudicate “private rights” must be vested in an Art. Ill court, see Part III, supra,

“this Court has accepted factfinding by an administrative agency,... as an adjunct to the Art. Ill court, analogizing the agency to a jury or a special master and permitting it in admiralty cases to perform the function of the special master. Crowell v. Benson, 285 U. S. 22, 51-*7865 (1932).” Atlas Roofing Co. v. Occupational Safety and Health Review Comm’n, 430 U. S., at 450, n. 7.

The use of administrative agencies as adjuncts was first upheld in Crowell v. Benson, supra. The congressional scheme challenged in Crowell empowered an administrative agency, the United States Employees’ Compensation Commission, to make initial factual determinations pursuant to a federal statute requiring employers to compensate their employees for work-related injuries occurring upon the navigable waters of the United States. The Court began its analysis by noting that the federal statute administered by the Compensation Commission provided for compensation of injured employees “irrespective of fault,” and that the statute also prescribed a fixed and mandatory schedule of compensation. Id., at 38. The agency was thus left with the limited role of determining “questions of fact as to the circumstances, nature, extent and consequences of the injuries sustained by the employee for which compensation is to be made.” Id., at 54. The agency did not possess the power to enforce any of its compensation orders: On the contrary, every compensation order was appealable to the appropriate federal district court, which had the sole power to enforce it or set it aside, depending upon whether the court determined it to be “in accordance with law” and supported by evidence in the record. Id., at 44-45, 48. The Court found that in view of these limitations upon the Compensation Commission’s functions and powers, its determinations were “closely analogous to findings of the amount of damages that are made, according to familiar practice, by commissioners or assessors.” Id., at 54. Observing that “there is no requirement that, in order to maintain the essential attributes of the judicial power, all determinations of fact in constitutional courts shall be made by judges,” id., at 51, the Court held that Art. Ill imposed no bar to the scheme enacted by Congress, id., at 54.

Crowell involved the adjudication of congressionally created rights. But this Court has sustained the use of adjunct factfinders even in the adjudication of constitutional rights— *79so long as those adjuncts were subject to sufficient control by an Art. Ill district court. In United States v. Raddatz, supra, the Court upheld the 1978 Federal Magistrates Act, which permitted district court judges to refer certain pretrial motions, including suppression motions based on alleged violations of constitutional rights, to a magistrate for initial determination. The Court observed that the magistrate’s proposed findings and recommendations were subject to de novo review by the district court, which was free to rehear the evidence or to call for additional evidence. Id., at 676-677, 681-683. Moreover, it was noted that the magistrate considered motions only upon reference from the district court, and that the magistrates were appointed, and subject to removal, by the district court. Id., at 685 (Blackmun, J., concurring).30 In short, the ultimate decisionmaking authority respecting all pretrial motions clearly remained with the district court. Id., at 682. Under these circumstances, the Court held that the Act did not violate the constraints of Art. III. Id., at 683-684.31

*80Together these cases establish two principles that aid us in determining the extent to which Congress may constitutionally vest traditionally judicial functions in non-Art. Ill officers. First, it is clear that when Congress creates a substantive federal right, it possesses substantial discretion to prescribe the manner in which that right may be adjudicated — including the assignment to an adjunct of some functions historically performed by judges.32 Thus Crowell rec*81ognized that Art. Ill does not require “all determinations of fact [to] be made by judges,” 285 U. S., at 51; with respect to congressionally created rights, some factual determinations may be made by a specialized factfinding tribunal designed by Congress, without constitutional bar, id., at 54. Second, the functions of the adjunct must be limited in such a way that “the essential attributes” of judicial power are retained in the Art. Ill court. Thus in upholding the adjunct scheme challenged in Crowell, the Court emphasized that “the reservation of full authority to the court to deal with matters of law provides for the appropriate exercise of the judicial function in this class of cases.” Ibid. And in refusing to invalidate the Magistrates Act at issue in Raddatz, the Court stressed that under the congressional scheme “ ‘[t]he authority — and the responsibility — to make an informed, final determination . . . remains with the judge,’” 447 U. S., at 682, quoting Mathews v. Weber, 423 U. S. 261, 271 (1976); the statute’s delegation of power was therefore permissible, since “the ultimate decision is made by the district court,” 447 U. S., at 683.

These two principles assist us in evaluating the “adjunct” scheme presented in these cases. Appellants assume that Congress’ power to create “adjuncts” to consider all cases related to those arising under Title 11 is as great as it was in the circumstances of Crowell. But while Crowell certainly endorsed the proposition that Congress possesses broad discretion to assign factfinding functions to an adjunct created to aid in the adjudication of congressionally created statutory rights, Crowell does not support the further proposition necessary to appellants’ argument — that Congress possesses the same degree of discretion in assigning traditionally judicial power to adjuncts engaged in the adjudication of rights not *82created by Congress. Indeed, the validity of this proposition was expressly denied in Crowell, when the Court rejected “the untenable assumption that the constitutional courts may be deprived in all cases of the determination of facts upon evidence even though a constitutional right may be involved,” 285 U. S., at 60-61 (emphasis added),33 and stated that

“the essential independence of the exercise of the judicial power of the United States in the enforcement of constitutional rights requires that the Federal court should determine ... an issue [of agency jurisdiction] upon its own record and the facts elicited before it.” Id., at 64 (emphasis added).34

Appellants’ proposition was also implicitly rejected in Raddatz. Congress’ assignment of adjunct functions under the Federal Magistrates Act was substantially narrower than under the statute challenged in Crowell. Yet the Court’s scrutiny of the adjunct scheme in Raddatz — which played a *83role in the adjudication of constitutional rights — was far stricter than it had been in Crowell. Critical to the Court’s decision to uphold the Magistrates Act was the fact that the ultimate decision was made by the district court. 447 U. S., at 683.

Although Crowell and Raddatz do not explicitly distinguish between rights created by Congress and other rights, such a distinction underlies in part Crowell’s and Raddatz’ recognition of a critical difference between rights created by federal statute and rights recognized by the Constitution. Moreover, such a distinction seems to us to be necessary in light of the delicate accommodations required by the principle of separation of powers reflected in Art. III. The constitutional system of checks and balances is designed to guard against “encroachment or aggrandizement” by Congress at the expense of the other branches of government. Buckley v. Valeo, 424 U. S., at 122. But when Congress creates a statutory right, it clearly has the discretion, in defining that right, to create presumptions, or assign burdens of proof, or prescribe remedies; it may also provide that persons seeking to vindicate that right must do so before particularized tribunals created to perform the specialized adjudicative tasks related to that right.35 Such provisions do, in a sense, affect the exercise of judicial power, but they are also incidental to Congress’ power to define the right that it has created. No *84comparable justification exists, however, when the right being adjudicated is not of congressional creation. In such a situation, substantial inroads into functions that have traditionally been performed by the Judiciary cannot be characterized merely as incidental extensions of Congress’ power to define rights that it has created. Rather, such inroads suggest unwarranted encroachments upon the judicial power of the United States, which our Constitution reserves for Art. Ill courts.

We hold that the Bankruptcy Act of 1978 carries the possibility of such an unwarranted encroachment. Many of the rights subject to adjudication by the Act’s bankruptcy courts, like the rights implicated in Raddatz, are not of Congress’ creation. Indeed, the cases before us, which center upon appellant Northern’s claim for damages for breach of contract and misrepresentation, involve a right created by state law, a right independent of and antecedent to the reorganization petition that conferred jurisdiction upon the Bankruptcy Court.36 Accordingly, Congress’ authority to control the manner in which that right is adjudicated, through assignment of historically judicial functions to a non-Art. Ill “adjunct,” plainly must be deemed at a minimum. Yet it is equally plain that Congress has vested the “adjunct” bankruptcy judges with powers over Northern’s state-created right that far exceed the powers that it has vested in administrative agencies that adjudicate only rights of Congress’ own creation.

Unlike the administrative scheme that we reviewed in Crowell, the Act vests all “essential attributes” of the judicial *85power of the United States in the “adjunct” bankruptcy-court. First, the agency in Crowell made only specialized, narrowly confined factual determinations regarding a particularized area of law. In contrast, the subject-matter jurisdiction of the bankruptcy courts encompasses not only traditional matters of bankruptcy, but also “all civil proceedings arising under title 11 or arising in or related to cases under title 11.” 28 U. S. C. § 1471(b) (1976 ed., Supp. IV) (emphasis added). Second, while the agency in Crowell engaged in statutorily channeled factfinding functions, the bankruptcy courts exercise “all of the jurisdiction” conferred by the Act on the district courts, § 1471(c) (emphasis added). Third, the agency in Crowell possessed only a limited power to issue compensation orders pursuant to specialized procedures, and its orders could be enforced only by order of the district court. By contrast, the bankruptcy courts exercise all ordinary powers of district courts, including the power to preside over jury trials, 28 U. S. C. § 1480 (1976 ed., Supp. IV), the power to issue declaratory judgments, § 2201, the power to issue writs of habeas corpus, § 2256, and the power to issue any order, process, or judgment appropriate for the enforcement of the provisions of Title 11, 11 U. S. C. § 105(a) (1976 ed., Supp. IV).37 Fourth, while orders issued by the agency in Crowell were to be set aside if “not supported by the evidence,” the judgments of the bankruptcy courts are apparently subject to review only under the more deferential “clearly erroneous” standard. See n. 5, supra. Finally, the agency in Crowell was required by law to seek enforcement of its compensation orders in the district court. In contrast, the bankruptcy courts issue final judgments, which are bind*86ing and enforceable even in the absence of an appeal.38 In short, the “adjunct” bankruptcy courts created by the Act exercise jurisdiction behind the facade of a grant to the district courts, and are exercising powers far greater than those lodged in the adjuncts approved in either Crowell or Raddatz.39

*87We conclude that 28 U. S. C. § 1471 (1976 ed., Supp. IV), as added by § 241(a) of the Bankruptcy Act of 1978, has impermissibly removed most, if not all, of “the essential attributes of the judicial power” from the Art. Ill district court, and has vested those attributes in a non-Art. Ill adjunct. Such a grant of jurisdiction cannot be sustained as an exercise of Congress’ power to create adjuncts to Art. Ill courts.

V

Having concluded that the broad grant of jurisdiction to the bankruptcy courts contained in 28 U. S. C. § 1471 (1976 ed., Supp. IV) is unconstitutional, we must now determine whether our holding should be applied retroactively to the effective date of the Act.40 Our decision in Chevron *88Oil Co. v. Huson, 404 U. S. 97 (1971), sets forth the three considerations recognized by our precedents as properly bearing upon the issue of retroactivity. They are, first, whether the holding in question “decid[ed] an issue of first impression whose resolution was not clearly foreshadowed” by earlier cases, id., at 106; second, “whether retrospective operation will further or retard [the] operation” of the holding in question, id., at 107; and third, whether retroactive application “could produce substantial inequitable results” in individual cases, ibid. In the present cases, all of these considerations militate against the retroactive application of our holding today. It is plain that Congress’ broad grant of judicial power to non-Art. Ill bankruptcy judges presents an unprecedented question of interpretation of Art. III. It is equally plain that retroactive application would not further the operation of our holding, and would surely visit substantial injustice and hardship upon those litigants who relied upon the Act’s vesting of jurisdiction in the bankruptcy courts. We hold, therefore, that our decision today shall apply only prospectively.41

The judgment of the District Court is affirmed. However, we stay our judgment until October 4, 1982. This limited stay will afford Congress an opportunity to reconstitute the bankruptcy courts or to adopt other valid means of adjudication, without impairing the interim administration of the bankruptcy laws. See Buckley v. Valeo, 424 U. S., at 143; *89cf. Georgia v. United States, 411 U. S. 526, 541 (1973); Fortson v. Morris, 385 U. S. 231, 235 (1966); Maryland Committee for Fair Representation v. Tawes, 377 U. S. 656, 675-676 (1964).

It is so ordered.

concurring in the judgment.

Were I to agree with the plurality that the question presented by these cases is “whether the assignment by Congress to bankruptcy judges of the jurisdiction granted in 28 U. S. C. §1471 (1976 ed., Supp IV) by § 241(a) of the Bankruptcy Act of 1978 violates Art. Ill of the Constitution,” ante, at 52, I would with considerable reluctance embark on the duty of deciding this broad question. But appellee Marathon Pipe Line Co. has not been subjected to the full range of authority granted bankruptcy courts by § 1471. It was named as a defendant in a suit brought by appellant Northern Pipeline Construction Co. in a United States Bankruptcy Court. The suit sought damages for, inter alia, breaches of contract and warranty. Marathon moved to dismiss the action on the grounds that the Bankruptcy Act of 1978, which authorized the suit, violated Art. Ill of the Constitution insofar as it established bankruptcy judges whose tenure and salary protection do not conform to the requirements of Art. III.

With the cases in this posture, Marathon has simply been named defendant in a lawsuit about a contract, a lawsuit initiated by appellant Northern after having previously filed a petition for reorganization under the Bankruptcy Act. Marathon may object to proceeding further with this lawsuit on the grounds that if it is to be resolved by an agency of the United States, it may be resolved only by an agency which exercises “[t]he judicial power of the United States” described by Art. Ill of the Constitution. But resolution of *90any objections it may make on this ground to the exercise of a different authority conferred on bankruptcy courts by the 1978 Act, see ante, at 54-55, should await the exercise of such authority.

“This Court, as is the case with all federal courts, ‘has no jurisdiction to pronounce any statute, either of a State or of the United States, void, because irreconcilable with the Constitution, except as it is called upon to adjudge the legal rights of litigants in actual controversies. In the exercise of that jurisdiction, it is bound by two rules, to which it has rigidly adhered, one, never to anticipate a question of constitutional law in advance of the necessity of deciding it; the other never to formulate a rule of constitutional law broader than is required by the precise facts to which it is to be applied.’ Liverpool, New York & Philadelphia S.S. Co. v. Commissioners of Emigration, 113 U. S, 33, 39.” United States v. Raines, 362 U. S. 17, 21 (1960).

Particularly in an area of constitutional law such as that of “Art. Ill Courts,” with its frequently arcane distinctions and confusing precedents, rigorous adherence to the principle that this Court should decide no more of a constitutional question than is absolutely necessary accords with both our decided cases and with sound judicial policy.

From the record before us, the lawsuit in which Marathon was named defendant seeks damages for breach of contract, misrepresentation, and other counts which are the stuff of the traditional actions at common law tried by the courts at Westminster in 1789. There is apparently no federal rule of decision provided for any of the issues in the lawsuit; the claims of Northern arise entirely under state law. No method of adjudication is hinted, other than the traditional common-law mode of judge and jury. The lawsuit is before the Bankruptcy Court only because the plaintiff has previously filed a petition for reorganization in that court.

*91The cases dealing with the authority of Congress to create courts other than by use of its power under Art. Ill do not admit of easy synthesis. In the interval of nearly 150 years between American Insurance Co. v. Canter, 1 Pet. 511 (1828), and Patmore v. United States, 411 U. S. 389 (1973), the Court addressed the question infrequently. I need not decide whether these cases in fact support a general proposition and three tidy exceptions, as the plurality believes, or whether instead they are but landmarks on a judicial “darkling plain” where ignorant armies have clashed by night, as Justice White apparently believes them to be. None of the cases has gone so far as to sanction the type of adjudication to which Marathon will be subjected against its will under the provisions of the 1978 Act. To whatever extent different powers granted under that Act might be sustained under the “public rights” doctrine of Murray’s Lessee v. Hoboken Land & Improvement Co., 18 How. 272 (1856), and succeeding cases, I am satisfied that the adjudication of Northern’s lawsuit cannot be so sustained.

I am likewise of the opinion that the extent of review by Art. Ill courts provided on appeal from a decision of the bankruptcy court in a case such as Northern’s does not save the grant of authority to the latter under the rule espoused in Crowell v. Benson, 285 U. S. 22 (1932). All matters of fact and law in whatever domains of the law to which the parties’ dispute may lead are to be resolved by the bankruptcy court in the first instance, with only traditional appellate review by Art. Ill courts apparently contemplated. Acting in this manner the bankruptcy court is not an “adjunct” of either the district court or the court of appeals.

I would, therefore, hold so much of the Bankruptcy Act of 1978 as enables a Bankruptcy Court to entertain and decide Northern’s lawsuit over Marathon’s objection to be violative of Art. Ill of the United States Constitution. Because I agree with the plurality that this grant of authority is not readily severable from the remaining grant of authority to *92bankruptcy courts under § 1471, see ante, at 87-88, n. 40, I concur in the judgment. I also agree with the discussion in Part V of the plurality opinion respecting retroactivity and the staying of the judgment of this Court.

dissenting.

I join Justice White’s dissenting opinion, but I write separately to emphasize that, notwithstanding the plurality opinion, the Court does not hold today that Congress’ broad grant of jurisdiction to the new bankruptcy courts is generally inconsistent with Art. Ill of the Constitution. Rather, the Court’s holding is limited to the proposition stated by Justice Rehnquist in his concurrence in the judgment — that a “traditional” state common-law action, not made subject to a federal rule of decision, and related only peripherally to an adjudication of bankruptcy under federal law, must, absent the consent of the litigants, be heard by an “Art. Ill court” if it is to be heard by any court or agency of the United States. This limited holding, of course, does not suggest that there is something inherently unconstitutional about the new bankruptcy courts; nor does it preclude such courts from adjudicating all but a relatively narrow category of claims “arising under” or “arising in or relatéd to cases under” the Bankruptcy Act.

It will not be necessary for Congress, in order to meet the requirements of the Court’s holding, to undertake a radical restructuring of the present system of bankruptcy adjudication. The problems arising from today’s judgment can be resolved simply by providing that ancillary common-law actions, such as the one involved in these cases, be routed to the United States district court of which the bankruptcy court is an adjunct.

dissenting.

Article III, § 1, of the Constitution is straightforward and uncomplicated on its face:

*93“The judicial Power of the United States, shall be vested in one supreme Court, and in such inferior Courts as the Congress may from time to time ordain and establish. The Judges, both of the supreme and inferior Courts, shall hold their Offices during good Behaviour, and shall at stated Times, receive for their Services, a Compensation, which shall not be diminished during their Continuance in Office.”

Any reader could easily take this provision to mean that although Congress was free to establish such lower courts as it saw fit, any court that it did establish would be an “inferior” court exercising “judicial Power of the United States” and so must be manned by judges possessing both life tenure and a guaranteed minimal income. This would be an eminently sensible reading and one that, as the plurality shows, is well founded in both the documentary sources and the political doctrine of separation of powers that stands behind much of our constitutional structure. Ante, at 57-60.

If this simple reading were correct and we were free to disregard 150 years of history, these would be easy cases and the plurality opinion could end with its observation that “[i]t is undisputed that the bankruptcy judges whose offices were created by the Bankruptcy Act of 1978 do not enjoy the protections constitutionally afforded to Art. Ill judges.” Ante, at 60. The fact that the plurality must go on to deal with what has been characterized as one of the most confusing and controversial areas of constitutional law1 itself indicates the gross oversimplification implicit in the plurality’s claim that “our Constitution unambiguously enunciates a fundamental principle — that the ‘judicial Power of the United States’ must be reposed in an independent Judiciary [and] provides clear institutional protections for that independence.” Ibid. While this is fine rhetoric, analytically it serves only to put a distracting and superficial gloss on a difficult question.

*94That question is what limits Art. Ill places on Congress’ ability to create adjudicative institutions designed to carry out federal policy established pursuant to the substantive authority given Congress elsewhere in the Constitution. Whether fortunate or unfortunate, at this point in the history of constitutional law that question can no longer be answered by looking only to the constitutional text. This Court’s cases construing that text must also be considered. In its attempt to pigeonhole these cases, the plurality does violence to their meaning and creates an artificial structure that itself lacks coherence.

I

There are, I believe, two separate grounds for today’s decision. First, non-Art. Ill judges, regardless of whether they are labeled “adjuncts” to Art. Ill courts or “Art. I judges,” may consider only controversies arising out of federal law. Because the immediate controversy in these cases — Northern Pipeline’s claim against Marathon — arises out of state law, it may only be adjudicated, within the federal system, by an Art. Ill court.2 Second, regardless of the source of law that governs the controversy, Congress is prohibited by Art. Ill from establishing Art. I courts, with three narrow exceptions. Adjudication of bankruptcy proceedings does not fall within any of these exceptions. I shall deal with the first of these contentions in this section.

The plurality concedes that Congress may provide for initial adjudications by Art. I courts or administrative judges of all rights and duties arising under otherwise valid federal laws. Ante, at 80. There is no apparent reason why this principle should not extend to matters arising in federal bankruptcy proceedings. The plurality attempts to escape the reach of prior decisions by contending that the bankrupt’s claim against Marathon arose under state law. Non-Article III *95judges, in its view, cannot be vested with authority to adjudicate such issues. It then proceeds to strike down 28 U. S. C. §1471 (1976 ed., Supp. IV) on this ground. For several reasons, the Court’s judgment is unsupportable.

First, clearly this ground alone cannot support the Court’s invalidation of § 1471 on its face. The plurality concedes that in adjudications and discharges in bankruptcy, “the restructuring of debtor-creditor relations, which is at the core of the federal bankruptcy power,” ante, at 71, and “the manner in which the rights of debtors and creditors are adjusted,” ante, at 84, n. 36, are matters of federal law. Under the plurality’s own interpretation of the cases, therefore, these matters could be heard and decided by Art. I judges. But because the bankruptcy judge is also given authority to hear a case like that of appellant Northern against Marathon, which the Court says is founded on state law, the Court holds that the section must be stricken down on its face. This is a grossly unwarranted emasculation of the scheme Congress has adopted. Even if the Court is correct that such a state-law claim cannot be heard by a bankruptcy judge, there is no basis for doing more than declaring the section unconstitutional as applied to the claim against Marathon, leaving the section otherwise intact. In that event, cases such as these would have to be heard by Art. Ill judges or by state courts — unless the defendant consents to suit before the bankruptcy judge — just as they were before the 1978 Act was adopted. But this would remove from the jurisdiction of the bankruptcy judge only a tiny fraction of the cases he is now empowered to adjudicate and would not otherwise limit his jurisdiction.3

*96Second, the distinction between claims based on state law and those based on federal law disregards the real character of bankruptcy proceedings. The routine in ordinary bankruptcy cases now, as it was before 1978, is to stay actions against the bankrupt, collect the bankrupt’s assets, require creditors to file claims or be forever barred, allow or disallow claims that are filed, adjudicate preferences and fraudulent transfers, and make pro rata distributions to creditors, who will be barred by the discharge from taking further actions against the bankrupt. The crucial point to be made is that in the ordinary bankruptcy proceeding the great bulk of creditor claims are claims that have accrued under state law prior to bankruptcy — claims for goods sold, wages, rent, utilities, and the like. “[T]he word debt as used by the Act is not confined to its technical common law meaning but. . . extends to liabilities arising out of breach of contract ... to torts . . . and to taxes owing to the United States or state or local governments.” 1 W. Collier, Bankruptcy ¶ 1.14, p. 88 (14th ed. 1976). Every such claim must be filed and its validity is sub*97ject to adjudication by the bankruptcy court. The existence and validity of such claims recurringly depend on state law. Hence, the bankruptcy judge is constantly enmeshed in state-law issues.

The new aspect of the Bankruptcy Act of 1978, in this regard, therefore, is not the extension of federal jurisdiction to state-law claims, but its extension to particular kinds of state-law claims, such as contract cases against third parties or disputes involving property in the possession of a third person.4 Prior to 1978, a claim of a bankrupt against a third party, such as the claim against Marathon in this case, was not within the jurisdiction of the bankruptcy judge. The old limits were based, of course, on the restrictions implicit within the concept of in rem jurisdiction; the new extension is based on the concept of in personam jurisdiction. “The bankruptcy court is given in personam jurisdiction as well as in rem jurisdiction to handle everything that arises in a bankruptcy case.” H. R. Rep. No. 95-595, p. 445 (1977). The difference between the new and old Acts, therefore, is not to be found in a distinction between state-law and federal-law matters; rather, it is in a distinction between in rem and in personam jurisdiction. The majority at no place explains why this distinction should have constitutional implications.

Third, all that can be left of the majority’s argument in this regard is that state-law claims adjudicated within the federal system must be heard in the first instance by Art. Ill judges. I shall argue below that any such attempt to distinguish Art. I from Art. Ill courts by the character of the controversies they may adjudicate fundamentally misunderstands the his*98torical and constitutional significance of Art. I courts. Initially, however, the majority’s proposal seems to turn the separation-of-powers doctrine, upon which the majority relies, on its head: Since state-law claims would ordinarily not be heard by Art. Ill judges — i. e., they would be heard by state judges — one would think that there is little danger of a diminution of, or intrusion upon, the power of Art. Ill courts, when such claims are assigned to a non-Art. Ill court. The plurality misses this obvious point because it concentrates on explaining how it is that federally created rights can ever be adjudicated in Art. I courts — a far more difficult problem under the separation-of-powers doctrine. The plurality fumbles when it assumes that the rationale it develops to deal with the latter problem must also govern the former problem. In fact, the two are simply unrelated and the majority never really explains the separation-of-powers problem that would be created by assigning state-law questions to legislative courts or to adjuncts of Art. Ill courts.

One need not contemplate the intricacies of the separation-of-powers doctrine, however, to realize that the majority’s position on adjudication of state-law claims is based on an abstract theory that has little to do with the reality of bankruptcy proceedings. Even prior to the present Act, bankruptcy cases were generally referred to bankruptcy judges, previously called referees. Bkrtcy. Rule 102(a). Title 11 U. S. C. §66 described the jurisdiction of the referees. Their powers included the authority to

“consider all petitions referred to them and make the adjudications or dismiss the petitions . . . grant, deny or revoke discharges, determine the dischargeability of debts, and render judgments thereon [and] perform such of the duties as are by this title conferred on courts of bankruptcy, including those incidental to ancillary jurisdiction, and as shall be prescribed by rules or orders of the courts of bankruptcy of their respective districts, except as herein otherwise provided.”

*99The bankruptcy judge possessed “complete jurisdiction of the proceedings.” 1 W. Collier, Bankruptcy ¶ 1.09, p. 65 (14th ed. 1976). The referee would initially hear and decide practically all matters arising in the proceedings, including the allowance and disallowance of the claims of creditors.5 If a claim was disallowed by the bankruptcy judge and the decision was not reversed on appeal, the creditor was forever barred from further action against the bankrupt. As pointed out above, all of these matters could and usually did involve state-law issues. Initial adjudication of state-law issues by non-Art. Ill judges is, then, hardly a new aspect of the 1978 Act.

Furthermore, I take it that the Court does not condemn as inconsistent with Art. Ill the assignment of these functions — i. e., those within the summary jurisdiction of the old bankruptcy courts — to a non-Art. Ill judge, since, as the plurality says, they lie at the core of the federal bankruptcy power. Ante, at 71. They also happen to be functions that have been performed by referees or bankruptcy judges for a very long time and without constitutional objection. Indeed, we approved the authority of the referee to allow or disallow claims in Katchen v. Landy, 382 U. S. 323 (1966). There, the referee held that a creditor had received a preference and that his claim could therefore not be allowed. We agreed that the referee had the authority not only to adjudicate the existence of the preference, but also to order that the preference be disgorged. We also recognized that the referee could adjudicate counterclaims against a creditor who files his claim against the estate. The 1973 Bankruptcy Rules make similar provision. See Rule 306(c), Rule 701, and Advisory Committee Note to Rule 701, 11 U. S. C., p. 1340. Hence, if Marathon had filed a claim against the bankrupt in this case, the trustee could have filed and the bankruptcy judge *100could have adjudicated a counterclaim seeking the relief that is involved in these cases.

Of course, all such adjudications by a bankruptcy judge or referee were subject to review in the district court, on the record. See 11 U. S. C. § 67(c). Bankruptcy Rule 810, transmitted to Congress by this Court, provided that the district court “shall accept the referee’s findings of fact unless they are clearly erroneous.” As the plurality recognizes, ante, at 55, the 1978 Act provides for appellate review in Art. Ill courts and presumably under the same “clearly-erroneous standard.” In other words, under both the old and new Acts, initial determinations of state-law questions were to be made by non-Art. Ill judges, subject to review by Art. Ill judges. Why the differences in the provisions for appeal in the two Acts are of unconstitutional dimension remains entirely unclear.

In theory and fact, therefore, I can find no basis for that part of the majority’s argument that rests on the state-law character of the claim involved here. Even if, prior to 1978, the referee could not generally participate in cases aimed at collecting the assets of a bankrupt estate, he nevertheless repeatedly adjudicated issues controlled by state law. There is very little reason to strike down § 1471 on its face on the ground that it extends, in a comparatively minimal way, the referees’ authority to deal with state-law questions. To do so is to lose all sense of proportion.

II

The plurality unpersuasively attempts to bolster its case for facial invalidity by asserting that the bankruptcy courts are now “exercising powers far greater than those lodged in the adjuncts approved in either Crowell or Raddatz.” Ante, at 86. In support of this proposition it makes five arguments in addition to the “state-law” issue. Preliminarily, I see no basis for according standing to Marathon to raise any of these additional points. The state-law objection applies to *101the Marathon case. Only that objection should now be adjudicated.6

I also believe that the major premise of the plurality’s argument is wholly unsupported: There is no explanation of why Crowell v. Benson, 285 U. S. 22 (1932), and United States v. Raddatz, 447 U. S. 667 (1980), define the outer limits of constitutional authority. Much more relevant to today’s decision are, first, the practice in bankruptcy prior to 1978, which neither the majority nor any authoritative case has questioned, and, second, the practice of today’s administrative agencies. Considered from this perspective, all of the plurality’s arguments are unsupportable abstractions, divorced from the realities of modern practice.

The first three arguments offered by the plurality, ante, at 85, focus on the narrowly defined task and authority of the agency considered in Crowell: The agency made only “specialized, narrowly confined factual determinations” and could issue only a narrow class of orders. Regardless of whether this was true of the Compensation Board at issue in Crowell, it certainly was not true of the old bankruptcy courts, nor does it even vaguely resemble current administrative practice. As I have already said, general references to bankruptcy judges, which was the usual practice prior to 1978, permitted bankruptcy judges to perform almost all of the functions of a bankruptcy court. Referees or bankruptcy judges not only exercised summary jurisdiction but could also conduct adversary proceedings to

“(1) recover money or property .... (2) determine the validity, priority, or extent of a lien or other interest in property, (3) sell property free of a lien or other interest for which the holder can be compelled to make a money satisfaction, (4) object to or revoke a discharge, (5) obtain an injunction, (6) obtain relief from a stay ... (7) determine the dischargeability of a debt.” Bkrtcy. Rule 701.

*102Although there were some exceptions to the referees’ authority, which have been removed by the 1978 Act, the additions to the jurisdiction of the bankruptcy judges were of marginal significance when examined in the light of the overall functions of those judges before and after 1978. In my view, those changes are not sufficient to work a qualitative change in the character of the bankruptcy judge.

The plurality’s fourth argument fails to point to any difference between the new and old Bankruptcy Acts. While the administrative orders in Crowell may have been set aside by a court if “not supported by the evidence,” under both the new and old Acts at issue here, orders of the bankruptcy judge are reviewed under the “clearly-erroneous standard.” See Bkrtcy. Rule 810. Indeed, judicial review of the orders of bankruptcy judges is more stringent than that of many modern administrative agencies. Generally courts are not free to set aside the findings of administrative agencies, if supported by substantial evidence. But more importantly, courts are also admonished to give substantial deference to the agency’s interpretation of the statute it is enforcing. No such deference is required with respect to decisions on the law made by bankruptcy judges.

Finally, the plurality suggests that, unlike the agency considered in Crowell, the orders of a post-1978 bankruptcy judge are final and binding even though not appealed. Ante, at 85-86. To attribute any constitutional significance to this, unless the plurality intends to throw into question a large body of administrative law, is strange. More directly, this simply does not represent any change in bankruptcy practice. It was hornbook law prior to 1978 that the authorized judgments and orders of referees, including turnover orders, were final and binding and res judicata unless appealed and overturned:

“The practice before the referee should not differ from that before the judge of the court of bankruptcy and, apart from direct review within the limitation of § 39(c), *103the orders of the referee are entitled to the same presumption of validity, conclusiveness and recognition in the court of bankruptcy or other courts.” 1 W. Collier, Bankruptcy ¶ 1.09, pp. 65-66 (14th ed. 1976).

Even if there are specific powers now vested in bankruptcy judges that should be performed by Art. Ill judges, the great bulk of their functions are unexceptionable and should be left intact. Whatever is invalid should be declared to be such; the rest of the 1978 Act should be left alone. I can account for the majority’s inexplicably heavy hand in this case only by assuming that the Court has once again lost its conceptual bearings when confronted with the difficult problem of the nature and role of Art. I courts. To that question I now turn.

Ill

A

The plurality contends that the precedents upholding Art. I courts can be reduced to three categories. First, there are territorial courts, which need not satisfy Art. Ill constraints because “the Framers intended that as to certain geographical areas . . . Congress was to exercise the general powers of government.”7 Ante, at 64. Second, there are courts-martial, which are exempt from Art. Ill limits because of a constitutional grant of power that has been “historically understood as giving the political Branches of Government extraordinary control over the precise subject matter at issue.” Ante, at 66. Finally, there are those legislative courts and administrative agencies that adjudicate cases involving public rights — controversies between the Government and private parties — which are not covered by Art. Ill because the controversy could have been resolved by the ex*104ecutive alone without judicial review. See ante, at 68. Despite the plurality’s attempt to cabin the domain of Art. I courts, it is quite unrealistic to consider these to be only three “narrow,” ante, at 64, limitations on or exceptions to the reach of Art. III. In fact, the plurality itself breaks the mold in its discussion of “adjuncts” in Part IV, when it announces that “when Congress creates a substantive federal right, it possesses substantial discretion to prescribe the manner in which that right may be adjudicated.” Ante, at 80. Adjudications of federal rights may, according to the plurality, be committed to administrative agencies, as long as provision is made for judicial review.

The first principle introduced by the plurality is geographical: Art. I courts presumably are not permitted within the States.8 The problem, of course, is that both of the other exceptions recognize that Art. I courts can indeed operate within the States. The second category relies upon a new principle: Art. I courts are permissible in areas in which the Constitution grants Congress “extraordinary control over the precise subject matter.” Ante, at 66. Preliminarily, I do not know how we are to distinguish those areas in which Congress’ control is “extraordinary” from those in which it is not. Congress’ power over the Armed Forces is established in Art. I, § 8, els. 13, 14. There is nothing in those Clauses that creates congressional authority different in kind from the authority granted to legislate with respect to bankruptcy. But more importantly, in its third category, and in its treatment of “adjuncts,” the plurality itself recognizes that Congress can create Art. I courts in virtually all the areas in which Congress is authorized to act, regardless of the quality of the constitutional grant of authority. At the same time, *105territorial courts or the courts of the District of Columbia, which are Art. I courts, adjudicate private, just as much as public or federal, rights.

Instead of telling us what it is Art. I courts can and cannot do, the plurality presents us with a list of Art. I courts. When we try to distinguish those courts from their Art. Ill counterparts, we find — apart from the obvious lack of Art. Ill judges — a series of nondistinctions. By the plurality’s own admission, Art. I courts can operate throughout the country, they can adjudicate both private and public rights, and they can adjudicate matters arising from congressional actions in those areas in which congressional control is “extraordinary.” I cannot distinguish this last category from the general “arising under” jurisdiction of Art. Ill courts.

The plurality opinion has the appearance of limiting Art. I courts only because it fails to add together the sum of its parts. Rather than limiting each other, the principles relied upon complement each other; together they cover virtually the whole domain of possible areas of adjudication. Without a unifying principle, the plurality’s argument reduces to the proposition that because bankruptcy courts are not sufficiently like any of these three exceptions, they may not be either Art. I courts or adjuncts to Art. Ill courts. But we need to know why bankruptcy courts cannot qualify as Art. I courts in their own right.

B

The plurality opinion is not the first unsuccessful attempt to articulate a principled ground by which to distinguish Art. I from Art. Ill courts. The concept of a legislative, or Art. I, court was introduced by an opinion authored by Chief Justice Marshall. Not only did he create the concept, but at the same time he started the theoretical controversy that has ever since surrounded the concept:

“The Judges of the Superior Courts of Florida hold their offices for four years. These Courts, then, are not constitutional Courts, in which the judicial power conferred *106by the Constitution on the general government, can be deposited. They are incapable of receiving it. They are legislative Courts, created in virtue of the general right of sovereignty which exists in the government, or in virtue of that clause which enables Congress to make all needful rules and regulations, respecting the territory belonging to the United States. The jurisdiction with which they are invested, is not a part of that judicial power which is defined in the 3d article of the Constitution, but is conferred by Congress, in the execution of those general powers which that body possesses over the territories of the United States.” American Insurance Co. v. Canter, 1 Pet. 511, 546 (1828).

The proposition was simple enough: Constitutional courts exercise the judicial power described in Art. Ill of the Constitution; legislative courts do not and cannot.

There were only two problems with this proposition. First, Canter itself involved a case in admiralty jurisdiction, which is specifically included within the “judicial power of the United States” delineated in Art. III. How, then, could the territorial court not be exercising Art. Ill judicial power? Second, and no less troubling, if the territorial courts could not exercise Art. Ill power, how could their decisions be subject to appellate review in Art. Ill courts, including this one, that can exercise only Art. Ill “judicial” power? Yet from early on this Court has exercised such appellate jurisdiction. Benner v. Porter, 9 How. 235, 243 (1850); Clinton v. Englebrecht, 13 Wall. 434 (1872); Reynolds v. United States, 98 U. S. 145, 154 (1879); United States v. Coe, 155 U. S. 76, 86 (1894); Balzac v. Porto Rico, 258 U. S. 298, 312-313 (1922). The attempt to understand the seemingly unexplainable was bound to generate “confusion and controversy.” This analytic framework, however — the search for a principled distinction — has continued to burden the Court.

The first major elaboration on the Canter principle was in Murray’s Lessee v. Hoboken Land & Improvement Co., 18 *107How. 272 (1856). The plaintiff in that case argued that a proceeding against a customs collector for the collection of moneys claimed to be due to the United States was an exercise of “judicial power” and therefore had to be carried out by Art. Ill judges. The Court accepted this premise: “It must be admitted that, if the auditing of this account, and the ascertainment of its balance, and the issuing of this process, was an exercise of the judicial power of the United States, the proceeding was void; for the officers who performed these acts could exercise no part of that judicial power.” Id., at 275. Having accepted this premise, the Court went on to delineate those matters which could be determined only by an Art. Ill court, i. e., those matters that fall within the nondelegable “judicial power” of the United States. The Court’s response to this was twofold. First, it suggested that there are certain matters which are inherently “judicial”: “[W]e do not consider congress can either withdraw from judicial cognizance any matter which, from its nature, is the subject of a suit at the common law, or in equity, or admiralty.” Id., at 284. Second, it suggested that there is another class of issues that, depending upon the form in which Congress structures the decisionmaking process, may or may not fall within “the cognizance of the courts of the United States.” Ibid. This latter category consisted of the so-called “public rights.” Apparently, the idea was that Congress was free to structure the adjudication of “public rights” without regard to Art. III.

Having accepted the plaintiff’s premise, it is hard to see how the Court could have taken too seriously its first contention. The Court presented no examples of such issues that are judicial “by nature” and simply failed to acknowledge that Art. I courts already sanctioned by the Court — e. g., territorial courts — were deciding such issues all the time. The second point, however, contains implicitly a critical insight; one that if openly acknowledged would have undermined the entire structure. That insight follows from the Court’s earlier *108recognition that the term “judicial act” is broad enough to encompass all administrative action involving inquiry into facts and the application of law to those facts. Id., at 280. If administrative action can be characterized as “judicial” in nature, then obviously the Court’s subsequent attempt to distinguish administrative from judicial action on the basis of the manner in which Congress structures the decision cannot succeed. There need be no Art. Ill court involvement in any adjudication of a “public right,” which the majority now interprets as any civil matter arising between the Federal Government and a citizen. In that area, whether an issue is to be decided by an Art. Ill court depends, finally, on congressional intent.

Although Murray’s Lessee implicitly undermined Chief Justice Marshall’s suggestion that there is a difference in kind between the work of Art. I and that of Art. Ill courts, it did not contend that the Court must always defer to congressional desire in this regard. The Court considered the plaintiff’s contention that removal of the issue from an Art. Ill court must be justified by “necessity.” Although not entirely clear, the Court seems to have accepted this proposition: “[I]t seems to us that the just inference from the entire law is, that there was such a necessity for the warrant.” Id., at 285.9

The Court in Murray’s Lessee was precisely right: Whether an issue can be decided by a non-Art. Ill court does not depend upon the judicial or nonjudicial character of the issue, but on the will of Congress and the reasons Congress offers for not using an Art. Ill court. This insight, however, was completely disavowed in the next major case to consider *109the distinction between Art. I and Art. Ill courts, Ex parte Bakelite Corp., 279 U. S. 438 (1929), in which the Court concluded that the Court of Customs Appeals was a legislative court. The Court there directly embraced the principle also articulated in Murray’s Lessee that Art. I courts may not consider any matter “which inherently or necessarily requires judicial determination,” but only such matters as are “susceptible of legislative or executive determination.” 279 U. S., at 453. It then went on effectively to bury the critical insight of Murray’s Lessee, labeling as “fallacious” any argument that “assumes that whether a court is of one class or the other depends on the intention of Congress, whereas the true test lies in the power under which the court was created and in the jurisdiction conferred.” 279 U. S., at 459.10

The distinction between public and private rights as the principle delineating the proper domains of legislative and constitutional courts respectively received its death blow, I had believed, in Crowell v. Benson, 285 U. S. 22 (1932). In that case, the Court approved an administrative scheme for the determination, in the first instance, of maritime employee compensation claims. Although acknowledging the framework set out in Murray’s Lessee and Ex parte Bakelite Corp., the Court specifically distinguished the case before it: “The present case does not fall within the categories just described but is one of private right, that is, of the liability of one individual to another under the law as defined.”11 285 U. S., at 51. Nevertheless, the Court approved of the use of an Art. I adjudication mechanism on the new theory that “there is no requirement that, in order to maintain the essen*110tial attributes of the judicial power, all determinations of fact in constitutional courts shall be made by judges.” Ibid. Article I courts could deal not only with public rights, but also, to an extent, with private rights. The Court now established a distinction between questions of fact and law: “[T]he reservation of full authority to the court to deal with matters of law provides for the appropriate exercise of the judicial function in this class of cases.”12 Id., at 54.

Whatever sense Crowell may have seemed to give to this subject was exceedingly short-lived. One year later, the Court returned to this subject, abandoning both the public/private and the fact/law distinction and replacing both with a simple literalism. In O’Donoghue v. United States, 289 U. S. 516 (1933), considering the courts of the District of Columbia, and in Williams v. United States, 289 U. S. 553 (1933), considering the Court of Claims, the Court adopted the principle that if a federal court exercises jurisdiction over cases of the type listed in Art. Ill, § 2, as falling within the “judicial power of the United States,” then that court must be an Art. Ill court:

“The provision of this section of the article is that the ‘judicial power shall extend’ to the cases enumerated, and it logically follows that where jurisdiction over these cases is conferred upon the courts of the District, the judicial power, since they are capable of receiving it, is ipso facto, vested in such courts as inferior courts of the United States.” O’Donoghue, supra, at 545.13

*111In order to apply this same principle and yet hold the Court of Claims to be a legislative court, the Court found it necessary in Williams, supra, to conclude that the phrase “Controversies to which the United States shall be a party” in Art. Ill must be read as if it said “Controversies to which the United States shall be a party plaintiff or petitioner.”14

By the time of the Williams decision, this area of the law was mystifying to say the least. What followed helped very little, if at all. In the next two major cases the Court could not agree internally on a majority position. In National Insurance Co. v. Tidewater Co., 337 U. S. 582 (1949), the Court upheld a statute giving federal district courts jurisdiction over suits between citizens of the District of Columbia and citizens of a State. A majority of the Court, however, rejected the plurality position that Congress had the authority to assign Art. I powers to Art. Ill courts, at least outside of the District of Columbia. Only Chief Justice Vinson in dissent reflected on the other side of this problem: whether Art. I courts could be assigned Art. Ill powers. He entirely disagreed with the conceptual basis for Williams and O’Donoghue, noting that to the extent that Art. I courts consider non-Art. Ill matters, appellate review by an Art. Ill court would be precluded. Or conversely, since appellate review is exercised by this Court over Art. I courts, Art. I courts must “exercise federal question jurisdiction.” 337 U. S., at 643. Having gone this far, the Chief Justice was confronted with the obvious question of whether in fact “the distinction between constitutional and legislative courts is meaningless.” Id., at 644. Although suggesting that out*112side of the Territories or the District of Columbia there maybe some limits on assignment to Art. I courts of matters that fall within Art. Ill jurisdiction — apart from federal-question jurisdiction — for the most part the Chief Justice ended up relying on the good will of Congress: “[W]e cannot impute to Congress an intent now or in the future to transfer jurisdiction from constitutional to legislative courts for the purpose of emasculating the former.” Ibid.

Another chapter in this somewhat dense history of a constitutional quandary was provided by Justice Harlan’s plurality opinion in Glidden Co. v. Zdanok, 370 U. S. 530 (1962), in which the Court, despite Bakelite and Williams — and relying on an Act of Congress enacted since those decisions— held the Court of Claims and the Court of Customs and Patent Appeals to be Art. Ill courts. Justice Harlan continued the process of intellectual repudiation begun by Chief Justice Vinson in Tidewater. First, it was clear to him that Chief Justice Marshall could not have meant what he said in Canter on the inability of Art. I courts to consider issues within the jursidiction of Art. Ill courts: “Far from being ‘incapable of receiving’ federal-question jurisdiction, the territorial courts have long exercised a jurisdiction commensurate in this regard with that of the regular federal courts and have been subjected to the appellate jurisdiction of this Court precisely because they do so.” 370 U. S., at 545, n. 13. Second, exceptions to the requirements of Art. Ill, he thought, have not been founded on any principled distinction between Art. I issues and Art. Ill issues; rather, a “confluence of practical considerations,” id., at 547, accounts for this Court’s sanctioning of Art. I courts:

“The touchstone of decision in all these cases has been the need to exercise the jurisdiction then and there and for a transitory period. Whether constitutional limitations on the exercise of judicial power have been held inapplicable has depended on the particular local setting, the practical necessities, and the possible alternatives.” Id., at 547-548.

*113Finally, recognizing that there is frequently no way to distinguish between Art. I and Art. Ill courts on the basis of the work they do, Justice Harlan suggested that the only way to tell them apart is to examine the “establishing legislation” to see if it complies with the requirements of Art. III. This, however, comes dangerously close to saying that Art. Ill courts are those with Art. Ill judges; Art. I courts are those without such judges. One hundred and fifty years of constitutional history, in other words, had led to a simple tautology.

IV

The complicated and contradictory history of the issue before us leads me to conclude that Chief Justice Vinson and Justice Harlan reached the correct conclusion: There is no difference in principle between the work that Congress may assign to an Art. I court and that which the Constitution assigns to Art. Ill courts. Unless we want to overrule a large number of our precedents upholding a variety of Art. I courts — not to speak of those Art. I courts that go by the contemporary name of “administrative agencies” — this conclusion is inevitable. It is too late to go back that far; too late to return to the simplicity of the principle pronounced in Art. Ill and defended so vigorously and persuasively by Hamilton in The Federalist Nos. 78-82.

To say that the Court has failed to articulate a principle by which we can test the constitutionality of a putative Art. I court, or that there is no such abstract principle, is not to say that this Court must always defer to the legislative decision to create Art. I, rather than Art. Ill, courts. Article III is not to be read out of the Constitution; rather, it should be read as expressing one value that must be balanced against competing constitutional values and legislative responsibilities. This Court retains the final word on how that balance is to be struck.

Despite the principled, although largely mistaken, rhetoric expanded by the Court in this area over the years, such a balancing approach stands behind many of the decisions up*114holding Art. I courts. Justice Harlan suggested as much in Glidden, although he needlessly limited his consideration to the “temporary” courts that Congress has had to set up on a variety of occasions. In each of these instances, this Court has implicitly concluded that the legislative interest in creating an adjudicative institution of temporary duration outweighed the values furthered by a strict adherence to Art. III. Besides the territorial courts approved in American Insurance Co. v. Canter, 1 Pet. 511 (1828), these courts have included the Court of Private Land Claims, United States v. Coe, 155 U. S. 76 (1894), the Choctaw and Chickasaw Citizenship Court, Stephens v. Cherokee Nation, 174 U. S. 445 (1899), and consular courts established in foreign countries, In re Ross, 140 U. S. 453 (1891). This same sort of “practical” judgment was voiced, even if not relied upon, in Crowell with respect to the Employees’ Compensation Claims Commission, which was not meant to be of limited duration: “[W]e are unable to find any constitutional obstacle to the action of the Congress in availing itself of a method shown by experience to be essential in order to apply its standards to the thousands of cases involved.” 285 U. S., at 54. And even in Murray’s Lessee, there was a discussion of the “necessity” of Congress’ adopting an approach that avoided adjudication in an Art. Ill court. 18 How., at 285.

This was precisely the approach taken to this problem in Palmore v. United States, 411 U. S. 389 (1973), which, contrary to the suggestion of the plurality, did not rest on any theory of territorial or geographical control. Ante, at 75-76. Rather, it rested on an evaluation of the strength of the legislative interest in pursuing in this manner one of its constitutionally assigned responsibilities — a responsibility not different in kind from numerous other legislative responsibilities. Thus, Palmore referred to the wide variety of Art. I courts, not just territorial courts. It is in this light that the critical statement of the case must be understood:

*115“[T]he requirements of Art. Ill, which are applicable where laws of national applicability and affairs of national concern are at stake, must in proper circumstances give way to accommodate plenary grants of power to Congress to legislate with respect to specialized areas having particularized needs and warranting distinctive treatment.” 411 U. S., at 407-408.

I do not suggest that the Court should simply look to the strength of the legislative interest and ask itself if that interest is more compelling than the values furthered by Art. III. The inquiry should, rather, focus equally on those Art. Ill values and ask whether and to what extent the legislative scheme accommodates them or, conversely, substantially undermines them. The burden on Art. Ill values should then be measured against the values Congress hopes to serve through the use of Art. I courts.

To be more concrete: Crowell, supra, suggests that the presence of appellate review by an Art. Ill court will go a long way toward insuring a proper separation of powers. Appellate review of the decisions of legislative courts, like appellate review of state-court decisions, provides a firm check on the ability of the political institutions of government to ignore or transgress constitutional limits on their own authority. Obviously, therefore, a scheme of Art. I courts that provides for appellate review by Art. Ill courts should be substantially less controversial than a legislative attempt entirely to avoid judicial review in a constitutional court.

Similarly, as long as the proposed Art. I courts are designed to deal with issues likely to be of little interest to the political branches, there is less reason to fear that such courts represent a dangerous accumulation of power in one of the political branches of government. Chief Justice Vinson suggested as much when he stated that the Court should guard against any congressional attempt “to transfer jurisdiction *116. . . for the purpose of emasculating” constitutional courts. National Insurance Co. v. Tidewater Co., 337 U. S., at 644.

V

I believe that the new bankruptcy courts established by the Bankruptcy Act of 1978, 28 U. S. C. §1471 (1976 ed., Supp. IV), satisfy this standard.

First, ample provision is made for appellate review by Art. Ill courts. Appeals may in some circumstances be brought directly to the district courts. 28 U. S. C. § 1334 (1976 ed., Supp. IV). Decisions of the district courts are further appealable to the court of appeals. §1293. In other circumstances, appeals go first to a panel of bankruptcy judges, § 1482, and then to the court of appeals. § 1293. In still other circumstances — when the parties agree — appeals may go directly to the court of appeals. In sum, there is in every instance a right of appeal to at least one Art. Ill court. Had Congress decided to assign all bankruptcy matters to the state courts, a power it clearly possesses, no greater review in an Art. Ill court would exist. Although I do not suggest that this analogy means that Congress may establish an Art. I court wherever it could have chosen to rely upon the state courts, it does suggest that the critical function of judicial review is being met in a manner that the Constitution suggests is sufficient.

Second, no one seriously argues that the Bankruptcy Act of 1978 represents an attempt by the political branches of government to aggrandize themselves at the expense of the third branch or an attempt to undermine the authority of constitutional courts in general. Indeed, the congressional perception of a lack of judicial interest in bankruptcy matters was one of the factors that led to the establishment of the bankruptcy courts: Congress feared that this lack of interest would lead to a failure by federal district courts to deal with bankruptcy matters in an expeditious manner. H. R. Rep. No. 95-595, p. 14 (1977). Bankruptcy matters are, for the most part, private adjudications of little political significance. *117Although some bankruptcies may indeed present politically controversial circumstances or issues, Congress has far more direct ways to involve itself in such matters than through some sort of subtle, or not so subtle, influence on bankruptcy judges. Furthermore, were such circumstances to arise, the Due Process Clause might very well require that the matter be considered by an Art. Ill judge: Bankruptcy proceedings remain, after all, subject to all of the strictures of that constitutional provision.15

Finally, I have no doubt that the ends that Congress sought to accomplish by creating a system of non-Art. Ill bankruptcy courts were at least as compelling as the ends found to be satisfactory in Palmore v. United States, 411 U. S. 389 (1973), or the ends that have traditionally justified the creation of legislative courts. The stresses placed upon the old bankruptcy system by the tremendous increase in bankruptcy cases were well documented and were clearly a matter to which Congress could respond.16 I do not believe it is possible to challenge Congress’ further determination that it was necessary to create a specialized court to deal with bankruptcy matters. This was the nearly uniform conclusion of all those that testified before Congress on the question of reform of the bankruptcy system, as well as the conclusion of the Commission on Bankruptcy Laws established by Congress in 1970 to explore possible improvements in the system.17

The real question is not whether Congress was justified *118in establishing a specialized bankruptcy court, but rather whether it was justified in failing to create a specialized, Art. Ill bankruptcy court. My own view is that the very fact of extreme specialization may be enough, and certainly has been enough in the past,18 to justify the creation of a legislative court. Congress may legitimately consider the effect on the federal judiciary of the addition of several hundred specialized judges: We are, on the whole, a body of generalists.19 The addition of several hundred specialists may substantially change, whether for good or bad, the character of the federal bench. Moreover, Congress may have desired to maintain some flexibility in its possible future responses to the general problem of bankruptcy. There is no question that the existence of several hundred bankruptcy judges with life tenure would have severely limited Congress’ future options. Furthermore, the number of bankruptcies may fluctuate, producing a substantially reduced need for bankruptcy judges. Congress may have thought that, in that event, a bankruptcy specialist should not as a general matter serve as a judge in the countless nonspecialized cases that come before the federal district courts. It would then face the prospect of large numbers of idle federal judges. Finally, Congress may have believed that the change from bankruptcy referees to Art. I judges was far less dramatic, and so less disruptive of the existing bankruptcy and constitutional court systems, than would be a change to Art. Ill judges.

For all of these reasons, I would defer to the congressional judgment. Accordingly, I dissent.