dissenting.
In my view, the Court’s holding is a grave repudiation of nearly 200 years of judicial precedent and historical practice. That repudiation arises out of an Act of Congress passed in the aftermath of a great national crisis which culminated in the resignation of a President. The Act (Title I of Pub. L. 93-526) violates firmly established constitutional principles in several respects.
*505I find it very disturbing that fundamental principles of constitutional law are subordinated to what seem the needs of a particular situation. That moments of great national distress give rise to passions reminds us why the three branches of Government were created as separate and coequal, each intended as a check, in turn, on possible excesses by one or both of the others. The Court, however, has now joined a Congress, in haste to “do something,” and has invaded historic, fundamental principles of the separate powers of coequal branches of Government. To “punish” one person, Congress— and now the Court — tears into the fabric of our constitutional framework.
Any case in this Court calling upon principles of separation of powers, rights of privacy, and the prohibitions against bills of attainder, whether urged by a former President — or any citizen — is inevitably a major constitutional holding. Mr. Justice Holmes, speaking of the tendency of “[g]reat cases like hard cases [to make] bad law,” went on to observe the dangers inherent when
“some accident of immediate overwhelming interest . . . appeals to the feelings and distorts the judgment. These immediate interests exercise a kind of hydraulic pressure which makes what previously was clear seem doubtful, and before which even well settled principles of law will bend.” Northern Securities Co. v. United States, 193 U. S. 197, 400-401 (1904) (dissenting opinion).
Well-settled principles of law are bent today by the Court under that kind of “hydraulic pressure.”
I
Separation of Powers
Appellant urges that Title I is an unconstitutional intrusion by Congress into the internal workings of the Office of the President, in violation of the constitutional principles of separation of powers. Three reasons support that conclusion. *506The well-established principles of separation of powers, as developed in the decisions of this Court, are violated if Congress compels or coerces the President, in matters relating to the operation and conduct of his office.1 Next, the Act is an exercise of executive — not legislative — power by the Legislative Branch. Finally, Title I works a sweeping modification of the constitutional privilege and historical practice of confidentiality of every Chief Executive since 1789.
A
As a threshold matter, we should first establish the standard of constitutional review by which Title I is to be judged. In the usual case, of course, legislation challenged in this Court benefits from a presumption of constitutionality. To survive judicial scrutiny a statutory enactment need only have a reasonable relationship to the promotion of an objective which the Constitution does not independently forbid, unless the legislation trenches on fundamental constitutional rights.
But where challenged legislation implicates fundamental constitutional guarantees, a far more demanding scrutiny is required. For example, this Court has held that the presumption of constitutionality does not apply with equal force where the very legitimacy of the composition of representative institutions is at stake. Reynolds v. Sims, 377 U. S. 533 (1964). Similarly, the presumption of constitutionality is lessened when the Court reviews legislation endangering fundamental constitutional rights, such as freedom of speech, or denying persons governmental rights or benefits because of race. Legislation touching substantially on these areas comes here bearing a heavy burden which its proponents must carry.
Long ago, this Court found the ordinary presumption of constitutionality inappropriate in measuring legislation directly impinging on the basic tripartite structure of our Government. *507In Kilbourn v. Thompson, 103 U. S. 168, 192 (1881), Mr. Justice Miller observed for the Court that encroachments by Congress posed the greatest threat to the continued independence of the other branches.2 Accordingly, he cautioned that the exercise of power by one branch directly affecting the potential independence of another “should be watched with vigilance, and when called in question before any other tribunal . . . should receive the most careful scrutiny.” Ibid. (Emphasis supplied.) See also Buckley v. Valeo, 424 U. S. 1 (1976).
Our role in reviewing legislation which touches on the fundamental structure of our Government is therefore akin to that which obtains when reviewing legislation touching on other fundamental constitutional guarantees. Because separation of powers is the base framework of our 'governmental system and the means by which all our liberties depend, Title I can be upheld only if it is necessary to secure some overriding governmental objective, and if there is no reasonable alternative which will trench less heavily on separation-of-powers principles.
B
Separation of powers is in no sense a formalism. It is the characteristic that distinguished our system from all others conceived up to the time of our Constitution. With federalism, separation of powers is “one of the two great structural principles of the American constitutional system . . . .” E. Corwin, The President 9 (1957). See also Griswold v. Connecticut, 381 U. S. 479, 501 (1965) (Harlan, J., concurring in judgment).
*508In pursuit of that principle, executive power was vested in the President; no other offices in the Executive Branch, other than the Presidency and Vice Presidency, were mandated by the Constitution. Only two Executive Branch offices, therefore, are creatures of the Constitution; all other departments and agencies, from the State Department to the General Services Administration, are creatures of the Congress and owe their very existence to the Legislative Branch.3
The Presidency, in contrast, stands on a very different footing. Unlike the vast array of departments which the President oversees, the Presidency is in no sense a creature of the Legislature. The President’s powers originate not from statute, but from the constitutional command to “take Care that the Laws be faithfully executed . . . .” These independent, constitutional origins of the Presidency have an important bearing on determining the appropriate extent of congressional power over the Chief Executive or his records and workpapers. For, although the branches of Government are obviously not divided into “watertight compartments,” Springer v. Philippine Islands, 277 U. S. 189, 211 (1928) (Holmes, J., dissenting), the office of the Presidency, as a constitutional equal of Congress, must as a general proposition be free from Congress’ coercive powers.4 This is not simply an abstract proposition *509of political philosophy; it is a fundamental prohibition plainly established by the decisions of this Court.
A unanimous Court, including Mr. Chief Justice Taft, Mr. Justice Holmes, and Mr. Justice Brandéis stated:
“The general rule is that neither department {of Government] may . . . control, direct or restrain the action of the other/’ Massachusetts v. Mellon, 262 U. S. 447, 488 (1923).
Similarly, in O’Donoghue v. United States, 289 U. S. 516, 530 (1933), the Court emphasized the need for each branch of Government to be free from the coercive influence of the other branches:
“[E]ach department should be kept completely independent of the others — independent not in the sense that they shall not cooperate to the common end of carrying into effect the purposes of the Constitution, but in the sense that the acts of each shall never be controlled by, or subjected, directly or indirectly, to, the coercive influence of either of the other departments.”
In Humphrey’s Executor v. United States, 295 U. S. 602, 629-630 (1935), the Court again held:
“The fundamental necessity of maintaining each of the three general departments of government entirely free from the control or coercive influence, direct or indirect, of either of the others, has often been • stressed and is hardly open to serious question. So much is implied in the very fact of the separation of the powers . . . .” (Emphasis supplied.)
Consistent with the principle of noncoercion, the unbroken practice since George Washington with respect to congressional demands for White House papers has been, in Mr. Chief Justice Taft’s words, that “while either house [of Congress] *510may request information, it cannot compel it . . . W. Taft, The Presidency 110 (1916). President Washington established the tradition by declining to produce papers requested by the House of Representatives relating to matters of foreign policy:
“To admit, then, a right in the House of Representatives to demand and to have as a matter of course all the papers respecting a negotiation with a foreign power would be to establish a dangerous precedent.” 1 Messages and Papers of the Presidents 195 (J. Richardson comp., 1899).
In noting the first President’s practice, this Court stated in United States v. Curtiss-Wright Corp., 299 U. S. 304, 320 (1936), that Washington’s historic precedent was “a refusal the wisdom of which was recognized by the House itself and has never since been doubted.” 5
Part of our constitutional fabric, then, from the beginning has been the President’s freedom from control or coercion by Congress, including attempts to procure documents that, though clearly pertaining to matters of important governmental interests, belong and pertain to the President. This freedom from Congress’ coercive influence, in the words of Humphrey’s Executor, “is implied in the very fact of the separation of the powers . . . .” 295 U. S., at 629-630. Moreover, it is not constitutionally significant that Congress has not directed that the papers be turned over to it for examination or retention, rather than to GSA. Separation of powers is fully implicated simply by Congress’ mandating what disposition is to be made of the papers of another branch.
This independence of the three branches of Government, including control over the papers of each, lies at the heart of *511this Court’s broad holdings concerning the immunity of congressional papers from outside scrutiny. The Constitution, of course, expressly grants immunity to Members of Congress as to any “Speech or Debate in either House . . yet the Court has refused to confine that Clause literally “to words spoken in debate.” Powell v. McCormack, 395 U. S. 486, 502 (1969). Congressional papers, including congressional reports, have been held protected by the Clause in order “ To prevent intimidation [of legislators] by the executive and accountability before a possibly hostile judiciary.’ ” Ibid. In a word, to preserve the constitutionally rooted independence of each branch of Government, each branch must be able to control its own papers.
Title I is an unprecedented departure from the constitutional tradition of noncompulsion. The statute commands the head of a legislatively created department to take and maintain custody of appellant’s Presidential papers, including many purely personal papers wholly unrelated to any operations of the Government. Title I does not concern itself in any way with materials belonging to departments of the Executive Branch created and controlled by Congress.
The Court brushes aside the fundamental principle of noncompulsion, abandoning outright the careful, previously unchallenged holdings of this Court in Mellon, O’Donoghue and Humphrey’s Executor. In place of this firmly established doctrine,6 the Court substitutes, without analysis, an ill-defined *512“pragmatic, flexible approach.” Ante, at 442. Recasting, for the immediate purposes of this case, our narrow holding in United States v. Nixon, 418 U. S. 683 (1974), see infra, at 515-516, the Court distills separation-of-powers principles into a simplistic rule which requires a “potential for disruption” or an “unduly disruptive” intrusion, before a measure will be held to trench on Presidential powers.7
The Court’s approach patently ignores Buckley v. Valeo, where, only one year ago, we unanimously found a separation-of-powers violation without any allegation, much less a showing, of “undue disruption.” There, we held that Congress could not impinge, even to the modest extent of six appointments to the Federal Election Commission, on the appointing powers of the President. We reached this conclusion in the face of the fact that President Ford had signed the bill into law.8
*513But even taking the “undue disruption” test as postulated, the Court engages in a facile analysis, as Mr. Justice Rehnquist so well demonstrates. We are told, under the Court’s view, that no “undue disruption” arises because GSA officials have taken custody of appellant’s Presidential papers, and since, for the time being, only GSA and other Executive Branch officials will have access to them. Ante, at 443-444.
This analysis is superficial in the extreme. Separation-of-powers principles are no less eroded simply because Congress goes through a “minuet” of directing Executive Department employees, rather than the Secretary of the Senate or the Doorkeeper of the House, to possess and control Presidential papers. Whether there has been a violation of separation-of-powers principles depends, not on the identity of the custodians, but upon which branch has commanded the custodians to act. Here, Congress has given the command.
If separation-of-powers principles can be so easily evaded, then the constitutional separation is a sham.
Congress’ power to regulate Executive Department documents, as contrasted with Presidential papers, under such measures as the Freedom of Information Act, 5 U. S. C. § 552 (1970 ed. and Supp. V), does not bear on the question. No one challenges Congress’ power to provide for access to records of the Executive Departments which Congress itself created. But the Freedom of Information Act, the Privacy Act of 1974, and similar measures never contemplated mandatory production of Presidential papers. What is instructive, by contrast, is the nonmandatory, noncoercive manner in which Congress has previously legislated with respect to Presidential papers, by providing for Presidential libraries at the option of every *514former President. Title I, however, breaches the nonmanda-tory tradition that has long been a vital incident of separation of powers.
C
The statute, therefore, violates separation-of-powers principles because it exercises a coercive influence by another branch over the Presidency. The legislation is also invalid on another ground pertaining to separation of powers; it is an attempt by Congress to exercise powers vested exclusively in the President — the power to control files, records, and papers of the office, which are comparable to the internal workpapers of Members of the House and Senate.
The general principle as to this aspect of separation of powers was stated in Kilbourn v. Thompson:
“[E]ach [branch] shall by the law of its creation be limited to the exercise of the powers appropriate to its own department and no other.
“[A]s a general rule . . . the powers confided by the Constitution to one of these departments cannot be exercised by another.” 103 TJ. S., at 191.
Madison also expressed this:
“For this reason that Convention which passed the ordinance of government, laid its foundation on this basis, that the legislative, executive and judiciary departments should be separate and distinct, so that no person should exercise the powers of more than one of them at the same time.” The Federalist No. 48, p. 335 (J. Cooke ed. 1961) (quoting Jefferson).
In the 1975 Term, in the face of a holding by a Court of Appeals that the separation-of-powers challenge was merit-less, we unanimously invalidated an attempt by Congress to exercise appointing powers constitutionally vested in the Chief Executive. Buckley v. Valeo, 424 U. S., at 109-143.
*515The Constitution does not speak of Presidential papers, just as it does not speak of workpapers of Members of Congress or of judges.9 But there can be no room for doubt that, up to now, it has been the implied prerogative of the President — as of Members of Congress and of judges — to memorialize matters, establish filing systems, and provide unilaterally for disposition of his workpapers. Control of Presidential papers is, obviously, a natural and necessary incident of the broad discretion vested in the President in order for him to discharge his duties.10
To be sure, we recognized a narrowly limited exception to Presidential control of Presidential papers in United States v. Nixon, 418 U. S. 683 (1974). But that case permits compulsory judicial intrusions only when a vital constitutional function, i. e., the conduct of criminal proceedings, would be impaired and when the President makes no more “than a generalized claim of . . . public interest . . . ,” id., at 707, in maintaining complete control of papers and in preserving confidentiality. That case, in short, was essentially a conflict between the Judicial Branch and the President, where the effective functioning of both branches demanded an accommodation and where the prosecutorial and judicial demands upon the President were very narrowly restricted with great *516specificity “to a limited number of conversations. . .Moreover, the request for production there was limited to materials that might themselves contain evidence of criminal activity of persons then under investigation or indictment. Finally, the intrusion was carefully limited to an in camera, examination, under strict limits, by a single United States District Judge. That case does not stand for the proposition that the Judiciary is at liberty to order all papers of a President into custody of United States Marshals.11
United States v. Nixon, therefore, provides no authority for Congress’ mandatory regulation of Presidential papers simply “to promote the general Welfare” which, of course, is a generalized purpose. No showing has been made, nor could it, that Congress’ functions will be impaired by the former President’s being allowed to control his own Presidential papers.12 Without any threat whatever to its own functions, Congress has by this statute, as in Buckley v. Valeo, exercised authority entrusted to the Executive Branch.13
*517D
Finally, in my view, the Act violates principles of separation of powers by intruding into the confidentiality of Presidential communications protected by the constitutionally based doctrine of Presidential privilege. A unanimous Court in United States v. Nixon could not have been clearer in holding that the privilege guaranteeing confidentiality of such communications derives from the Constitution, subject to compelled disclosure only in narrowly limited circumstances:
“A President and those who assist him must be free to explore alternatives in the process of shaping policies and making decisions and to do so in a way many would be unwilling to express except privately. These are the considerations justifying a presumptive privilege for Presidential communications. The privilege is fundamental to the operation of Government and inextricably rooted in the separation of powers under the Constitution.” 418 U. S., at 708.
President Lyndon Johnson expressed the historic view of Presidential confidentiality in even stronger terms in a letter to the GSA Administrator: “[S]ince the President . . . is the recipient of many confidences from others, and since the inviolability of such confidence is essential to the functioning of the constitutional office of the Presidency, it will be necessary to withhold from public scrutiny certain papers and *518classes of papers for varying periods of time. Therefore . . . I hereby reserve the right to restrict the use and availability of any materials ... for such time as I, in my sole discretion, may . . . specify . . . .” Hearing before a Subcommittee of the House Committee on Government Operations, on H. J. Res. 632, 89th Cong., 1st Sess., 17 (1965).
As a constitutionally based prerogative, Presidential privilege inures to the President himself; it is personal in the same sense as the privilege against compelled self-incrimination. Presidential privilege would therefore be largely illusory unless it could be interposed by the President against the countless thousands of persons in the Executive Branch, and most certainly if the executive officials are acting, as this statute contemplates, at the command of a different branch of Government.14
This statute requires that persons not designated or approved by the former President will review all Presidential papers. Even if the Government agents, in culling through the materials, follow the “advisory” suggestions offered by the District Court, the fact remains that their function abrogates the Presidential privilege. Congress has, in essence, commanded them to review and catalog thousands of papers and recordings that are undoubtedly privileged. Given that fact, it is clear that the Presidential privilege of one occupant of that office will have been rendered a nullity.15
*519E
There remains another inquiry under the issue of separation of powers. Does the fact that the Act applies only to a former President, described as “a legitimate class of one,” ante, at 472, after he has left office, justify what would otherwise be unconstitutional if applicable to an incumbent President?
On the face of it, congressional regulation of the papers of a former President obviously will have less disruptive impact on the operations of an incumbent President than an effort at regulation or control over the same papers of an incumbent President. But this “remoteness” does not eliminate the separation-of-powers defects. First, the principle that a President must be free from coercion should apply to a former President, so long as Congress is inquiring or acting with respect to operations of the Government while the former President was in office.16
To the extent Congress is empowered to coerce a former President, every future President is at risk of denial of a large measure of the autonomy and independence contemplated by the Constitution and of the confidentiality attending it. Myers v. United States, 272 U. S. 52 (1926). Indeed, the President, if he is to have autonomy while in office, needs the assurance that Congress will not immediately be free to coerce him to open all his files and records and give an account of Presidential actions at the instant his successor is sworn in.17 Absent the validity of the expectation of *520privacy of such papers (save for a subpoena under United States v. Nixon), future Presidents and those they consult will be well advised to take into account the possibility that their most confidential correspondence, workpapers, and diaries may well be open to congressionally mandated review, with no time limit, should some political issue give rise to an interbranch conflict.
The Need for Confidentiality
The consequences of this development on what a President expresses to others in writing and orally are incalculable; perhaps even more crucial is the inhibiting impact on those to whom the President turns for information and for counsel, whether they are officials in the Government, business or labor leaders, or foreign diplomats and statesmen. I have little doubt that Title I — and the Court's opinion- — will be the subject of careful scrutiny and analysis in the foreign offices of other countries whose representatives speak to a President on matters they prefer not to put in writing, but which may be memorialized by a President or an aide. Similarly, Title I may well be a “ghost” at future White House conferences, with conferees choosing their words more cautiously because of the enlarged prospect of compelled disclosure to others. A unanimous Court carefully took this into account in United States v. Nixon:
“The expectation of a President to the confidentiality of his conversations and correspondence, like the claim of confidentiality of judicial deliberations, for example, has all the values to which we accord deference for the privacy of all citizens and, added to those values, is the necessity for protection of the public interest in candid, objective, and even blunt or harsh opinions in Presidential decisionmaking.” 418 U. S., at 708.
*521In this same vein, Mr. Justice Powell argues that President Carter’s representation to the Court through the Solicitor General that Title I enhances the efficiency of the Executive Branch is dispositive of appellant’s separation-of-powers claim. This deference to the views of one administration, expressed approximately 100 days after its inception, as to the permanent structure of our Government is not supported by precedent and conflicts with 188 years of, history. First, there is no principled basis for limiting this unique deference. If and when the one-House veto issue, for example, comes before us, are we to accept the opinion of the Department of Justice as to the effects of that legislative device on the Executive Branch’s operations? Second, if Title I is thus efficacious, why did the President who signed this bill into law decide to establish a Presidential library in Ann Arbor, Mich., rather than turn all of his Presidential materials over to GSA for screening and retention in Washington, D. C., where the materials would be readily accessible to officials of the Executive Branch? And why, suddenly, is Congress’ acquiescence in President Ford’s actions consistent with the supposed foundation of Title I?
Third, as pointed out by Mr. Justice Blackmun, ante, at 491: “Political realities often guide a President to a decision not to veto” or, indeed, a decision not to challenge in court the actions of Congress. See n. 18, infra. Finally, it is perhaps not inappropriate to note that, on occasion, Presidents disagree with their predecessors on issues of policy. Some have believed in “Congressional Government”; others adhered to expansive notions of Presidential power. It is, I respectfully submit, a unique idea that this Court accept as controlling the representations of any administration on a constitutional question going to the permanent structure of Government.
Title I is also objectionable on separation-of-powers grounds, despite its applicability only to a former President, because compelling the disposition of all of a former President’s papers *522is a legislative exercise of what have historically been regarded as executive powers. Presidential papers do not, after all, instantly lose their nature quadrennially at high noon on January 20. Moreover, under Title I it is now the Congress, not the incumbent President,18 that has decided what to do with all the papers of one entire administration.
Finally, the federal appellees concede that Presidential privilege, a vital incident of our separation-of-powers system, does not terminate instantly upon a President’s departure from office. They candidly acknowledge that “the privilege survives the individual President’s tenure,” Brief for Federal Appellees 33, because of the vital public interests underlying the privilege. This principle, as all parties concede, finds explicit support in history; former President Truman in 1953 refused to provide information to the Congress on matters occurring during his administration, advising Congress:
“It must be obvious to you that if the doctrine of separation of powers and the independence of the Presidency is to have any validity at all, it must be equally applicable to a President after his term of office has expired when he is sought to be examined with respect to any acts occurring while he [was] President.” 120 Cong. Rec. 33419 (1974). (Emphasis supplied.) 19
*523To ensure institutional integrity and confidentiality, Presidents and their advisers must have assurance, as do judges and Members of Congress, that their internal communications will not become subject to retroactive legislation mandating intrusions into matters as to which there was a well-founded expectation of privacy when the communications took place. Just as Mr. Truman rejected congressional efforts to inquire of him, after he left office, as to his activities while President, this Court has always assumed that the immunity conferred by the Speech or Debate Clause is available to a Member of Congress after he leaves office. United States v. Brewster, 408 U. S. 501 (1972). It would therefore be illogical to conclude that the President loses all immunity from legislative coercion as to his Presidential papers from the moment he leaves office.
The Court correctly concedes that a former President retains the Presidential privilege after leaving office, ante, at 448-449; but it then concludes that several considerations cut against recognition of the privilege as to one former President. Pirst, the Court places great emphasis on the fact that neither President Ford nor President Carter “supports appellant’s claim . . . .” Ante, at 449. The relevance of that fact is not immediately clear. The validity of one person’s constitutional privilege does not depend on whether some other holder of the same privilege supports his claim.20 The fact that an incumbent President has signed or supports a particular measure cannot defeat a former President’s claim of privilege. If the Court is correct today, it was wrong one year ago in Buckley v. Valeo, when we unanimously held that Presidential approval of the Federal Election Cam*524paign Act could not validate an unconstitutional invasion of Presidential appointing authority.
Second, the Court suggests that many of the papers are unprivileged. Of the great volume of pages, appellant estimated that he saw only about 200,000 items while he was President. Several points are relevant in this regard. We do not know how many pages the 200,000 items represent; the critical factor is that all papers are presumptively privileged. Regardless of the number of pages, the fact remains that the 200,000 items that the President personally reviewed or prepared while in office obviously have greater historical value than the mass of routine papers coming to the White House. Mountains of Government reports tucked away in Presidential files will not likely engage the interest of archivists or historians, since most such reports are not historically important and are, in any event, available elsewhere. Rather, archivists and historians will want to find and preserve the materials that reflect the President’s internal decisionmaking processes. Those are precisely the papers which will be subjected to the most intensive review and which have always been afforded absolute protection. The Court’s analytically void invocation of sheer numbers cannot mask the fact that the targets of the review are privileged papers, diaries, and conversations.
I agree that, under United States v. Nixon, the Presidential privilege is qualified. From that premise, however, the Court leaps to the conclusion that future regulations governing public access to the materials are sufficient to protect that qualified privilege. The Act does indeed provide for a number of safeguards before the public at large obtains access to the materials. See § 104 (a). But the Court cannot have it both ways. The opinion expressly recognizes again and again that public access is not now the issue. The constitutionality of a statute cannot rest on the presumed validity of regulations not yet issued; moreover, no regulations governing public access can remedy the statute’s basic flaw of *525permitting Congress to seize the confidential papers of a President.
P
In concluding that Title I on its face violates the principle of separation of powers, I do not address the issue whether some circumstances might justify legislation for the disposition of Presidential papers without the President’s consent. Here, nothing remotely like the particularised need we found in United States v. Nixon has been shown with respect to these Presidential papers. No one has suggested that Congress will find its own "core” functioning impaired by lack of the impounded papers, as we expressly found the judicial function would be impaired by lack of the material subpoenaed in United States v. Nixon.
I leave to another day the question whether, under exigent circumstances, a narrowly defined congressional demand for Presidential materials might be justified. But Title I fails to satisfy either the required narrowness demanded by United States v. Nixon or the requirement that the coequal powers of the Presidency not be injured by congressional legislation.
II
Privacy
The discussion of separation of powers concerns, of course, the structure of government, not the rights of the sole individual ostensibly affected by this legislation. But Title I touches not only upon the independence of a coordinate branch of government, it also affects, in the most direct way, the basic rights of one named individual. The statute provides, as we have seen, for governmental custody over — and review of — all of the former President’s written and recorded materials at the time he left office, including diary recordings and conversations in his private residences outside Washington, D. C. § 101 (a)(2).
The District Court was deeply troubled by this admittedly *526unprecedented intrusion. Its opinion candidly acknowledged that the personal-privacy claim was the “most troublesome” point raised by this unique statute.21 In addition to communications and memoranda reflecting the President's confidential deliberations, the District Court admitted that the materials subject to GSA review included highly personal communications.
“Among all of the papers and tape recordings falling within the Act, however, are some papers and materials containing extremely private communications between [Mr. Nixon] and, among others,, his wife, his daughters, his physician, lawyer, and clergyman, and his close friends, as well as personal diary dictabelts and his wife’s personal files. . . . Segregating those that are private from those that are not private requires rather comprehensive screening, and archivists entrusted with that duty will be required to read or listen to private communications.” 408 F. Supp. 321, 359 (DC 1976).
A
Given this admitted intrusion, the legislation before us must be subjected to the most searching kind of judicial scrutiny.22 Statutes that trench on fundamental liberties, like *527those affecting significantly the structure of our government, are not entitled to the same presumption of constitutionality we normally accord legislation. Moore v. East Cleveland, 431 U. S. 494, 499 (1977). The burden of justification is reversed; the burden rests upon government, not on the individual whose liberties are affected, to justify the measure. Abood v. Detroit Board of Education, 431 U. S. 209, 263-264 (1977) (Powell, J., concurring in judgment). We recently reaffirmed the standard of review in such cases as one of “exacting scrutiny.”
“We long have recognized that significant encroachments on First Amendment rights of the sort that compelled disclosure imposes cannot be justified by a mere showing of some legitimate governmental interest .... [W]e have required that the subordinating interests of the State must survive exacting scrutiny.” Buckley v. Valeo, 424 U. S., at 64.
B
Constitutional analysis must, of course, take fully into account the nature of the Government’s interests underlying challenged legislation. Once those interests are identified, we must then focus on the nature of the individual interests affected by the statute. Id., at 14r-15. Finally, we must decide whether the Government’s interests are of sufficient weight to subordinate the individual’s interests, and, if so, whether the Government has nonetheless employed unnecessarily broad means for achieving its purposes. Lamont v. Postmaster General, 381 U. S. 301, 310 (1965) (Brennan, J., concurring).
Two governmental interests are asserted as the justification for this statute: to ensure the general efficiency of the Execu*528tive Branch’s operations23 and to preserve historically significant papers and tape recordings for posterity.24 Both these purposes are legitimate and important. Yet, there was no serious suggestion by Congress that the operations of the Executive Branch would actually be impaired unless, contrary to nearly 200 years’ past practice, all Presidential papers of the one named incumbent were required by law to be impounded in the sole control of Government agents. The statute on its face, moreover, does not purport to address a particularized need, such as the need to secure Presidential papers concerning the Middle East, the SALT talks, or problems in Panama.25 Indeed, the congressionally perceived “need” is a *529far more “generalized need” than that rejected in United States v. Nixon by a unanimous Court.
As to the interest in preserving historical materials, there is nothing whatever in our national experience to suggest that existing mechanisms, such as the 20-year-old Presidential Libraries Act, were insufficient to achieve that purpose.26 In any event, the interest in preserving “historical materials” cannot justify seizing, without notice or hearing, private papers preliminary to a line-by-line examination by Government agents.
In contrast to Congress’ purposes underlying the statute, this Act intrudes significantly on two areas of traditional privacy interests of Presidents. One embraces Presidential papers relating to his decisions, development of policies, appointments, and communications in his role as leader of a political party; the other encompasses purely private matters of family, property, investments, diaries, and intimate conversations. Both interests are of the highest order, with perhaps some primacy for family papers.27 Cf. Moore v. East Cleveland, supra, at 499.
Title I thus touches directly on what Mr. Justice Powell once referred to as the “intimate areas of an individual’s personal affairs,” California Bankers Assn. v. Shultz, 416 U. S. *53021, 78 (1974) (concurring opinion). The papers in both of these areas — family and political decisionmaking — are of the most private nature, enjoying the highest status under our law. Mr. Justice Brennan recently put it this way: “Personal letters constitute an integral aspect of a person’s private enclave.” Fisher v. United States, 425 U. S. 391, 427 (1976) (concurring in judgment). An individual’s papers, he said, are “an extension of his person.” Id., at 420. Mr. Justice Marshall made the same point: “Diaries and personal letters that record only their author’s personal thoughts lie at the heart of our sense of privacy.” Couch v. United States, 409 U. S. 322, 350 (1973) (dissenting opinion). In discussing private papers, he referred even more emphatically to the “deeply held belief on the part of the Members of this Court throughout its history that there are certain documents no person ought to be compelled to produce at the Government’s request.” Fisher v. United States, supra, at 431-432 (emphasis supplied) (concurring in judgment). This echoes Lord Camden’s oft-quoted description of personal papers as a man’s “dearest property.” Boyd v. United States, 116 U. S. 616, 628 (1886).
One point emerges clearly: The papers here involve the most fundamental First and Fourth Amendment interests. Since the Act asserts exclusive Government custody over all papers of a former President, the Fourth Amendment’s prohibition against unreasonable searches and seizures is surely implicated.28 Indeed, where papers or books are the subject *531of a government intrusion, our cases uniformly hold that the Fourth Amendment prohibition against a general search requires that warrants contain descriptions reflecting “the most scrupulous exactitude . . . ,” Stanford v. Texas, 379 U. S. 476, 485 (1965). Those cases proscribe general language in a warrant — or a statute — of “indiscriminate sweep . . . .” Id., at 486. Title I, commanding seizure followed by permanent control of all materials having “historical or commemorative value,” evidences the “indiscriminate sweep” we have long denounced. This “broad broom” statute provides virtually no standard at all to guide the Government agents combing through the papers; the agents are left to roam at large through confidential materials, something to which no other President and no Member of Congress or of the Judicial Branch has been subjected.
The Court, while recognizing that Government agents will necessarily be reviewing the most private kinds of communications covering a period of five and one-half years, tells us that Stanford is inapposite. Several reasons are given. The Court suggests that, unlike the instant case, the seizure in Stanford included vast quantities of materials unrelated to any legitimate government objective; in addition, the Stanford intrusion constituted an invasion of the home in connection with a criminal investigation. That last consideration relied on by the Court can be disposed of quickly, for by its terms, just as in Stanford, Title I commands seizure and review of papers from appellant’s private residences within and outside Washington, D. C., § 101 (a), for the purpose, among others, of criminal proceedings brought by the Special Prosecutor, § 102 (b), and to make the materials available more broadly “for use in judicial proceedings.” § 104 (a) (2). Title I is not needed for this purpose, since a narrowly defined subpoena can accomplish those purposes under United States v. Nixon. Title I is in effect a “legislative warrant” reminiscent of the odious general warrants of the colonial era.
*532As to the Court’s first consideration, its “quantity” test is fallacious. The intrusion in Stanford was unlawful not because the State had an interest in only part of many items in Stanford’s home, but rather because the warrant failed to describe the objects of seizure with the “most scrupulous exactitude.” Stanford is not a “numbers” test, the protection of which vanishes if unprotected materials outnumber protected materials; it is, rather, a test designed to ensure that protected materials are not seized at all. Title I on its face commands that protected materials be seized wherever found— including the private residences mentioned — reviewed, and returned only if the Government agents decide that certain protected materials lack historical significance. The Act plainly accomplishes exactly what Stanford expressly forbids.
In addition to Fourth Amendment considerations, highly important First Amendment interests pervade all Presidential papers, since they include expressions of privately held views about politics, diplomacy, or people of all walks of life, within and outside this country. Appellant’s freedom of association is also implicated, since his recordings and papers will likely reveal much about his relationships with both individuals and organizations. In NAACP v. Alabama, 357 U. S. 449, 462 (1958), the Court said:
“This Court has recognized the vital relationship between freedom to associate and privacy in one’s associations.”
Accordingly, in passing on a statute compelling disclosure of political contributions, the Court, in Buckley v. Valeo, imposed the strict standard of “exacting scrutiny” because of the significant impact on First Amendment rights.
The fact that the former President was an important national and world political figure obviously does not diminish the traditional privacy interest in his papers. Forced disclosure of private information, even to Government officials, is by no means sanctioned by this Court’s decisions, except for *533the most compelling reasons. Cf. Whalen v. Roe, 429 U. S. 589 (1977). I do not think, for example, that this Court would readily sustain, as a condition to holding public office, a requirement that a candidate reveal publicly membership in every organization whether religious, social, or political. After all, our decision in NAACP v. Alabama, supra, was presumably intended to protect from compelled disclosure members of the organization who were actively involved in public affairs or who held public office in Alabama.
The Court’s reliance on Whalen v. Roe, supra, in rejecting appellant’s privacy claim is surprising. That case dealt with the State’s undoubted police power to regulate dispensing of dangerous drugs, the very use or possession of which the State could forbid. 429 U. S., at 603, and 597 n. 20. Hence, we had no difficulty whatever in reaching a unanimous holding that the public interest in regulating dangerous drugs outweighed any privacy interest in reporting to the State all prescriptions, those reports being made confidential by statute. No personal, private business, or political confidences were involved.
C
In short, a former President up to now has had essentially the same expectation of privacy with respect to his papers and records as every other person. This expectation is soundly based on two factors: first, under our constitutional traditions, Presidential papers have been, for more than 180 years, deemed by the Congress to belong to the President. Congress ratified this tradition by specific Acts: (a) congressional appropriations following authorization to purchase Presidential papers; (b) congressional enactment of a non-mandatory system of Presidential libraries; and (c) statutes permitting, until 1969, a charitable-contribution deduction for papers of Presidents donated to the United States or to nonprofit institutions.
*534Second, in the absence of any legislation to the contrary, there was no reason whatever for a President to take time from his official duties to ensure that there was no “commingling” of “public” and “private” papers. Indeed, the fact that the former President commingled Presidential, and private family papers, absent any then-existing laws to the contrary, points strongly to the conclusion that he did in fact have an expectation of privacy with respect to both categories of papers.
On the basis of this Court’s holdings, I cannot understand why the former President’s privacy interests do not outweigh the generalized, undifferentiated goals sought to be achieved by Title I. Without a more carefully defined focus, these legislative goals do not represent “paramount Government interests,” nor is this particular piece of legislation needed to achieve those goals, even if we assume, arguendo, that they are of a “compelling” or “overriding” nature. But even if other Members of the Court strike the balance differently, the Government has nonetheless failed to choose narrowly tailored means of carrying out its purposes so as not unnecessarily to invade important First and Fourth Amendment liberties. The Court demanded no less in Buckley v. Valeo, and nothing less will do here. Cf. Hynes v. Mayor of Oradell, 425 U. S. 610, 620 (1976).
The federal appellees point to two factors as mitigating the effects of this admitted intrusion: first, in their view, most of the President’s papers and conversations relate to the business of Government, rather than to personal, family, or political matters; second, it is said that the intrusion is limited as much as possible, since the review will be carried out by specially trained Government agents.
Even accepting the Government’s interest in identifying and preserving govemmentally related papers in order to preserve them for historical purposes, that interest cannot justify a seizure and search of dll the papers taken here. *535Since compulsory review of personal and family papers and tape recordings is an admittedly improper invasion of privacy, no constitutional principle justifies an intrusion into indisputably protected areas in order to carry out the “generalized” statutory objectives.
Second, the intrusion cannot be saved by the credentials, however impeccable, of the Government agents. The initial problem with this justification is that no one knows whether these agents are, as the federal appellees contend, uniformly discreet. Despite the lip service paid by the District Court and appellees to the record of archivists generally, there is nothing before us to justify the conclusion that each of the more than 100 persons who apparently will have access to, and will monitor and examine, the materials is indeed reliably discreet.
The Act, furthermore, provides GSA with no meaningful standards to minimize the extent of intrusions upon appellant’s privacy. We are thus faced with precisely the same standardless discretion vested in governmental officials which this Court has unhesitatingly struck down in other First Amendment areas. See, e. g., Hynes v. Mayor of Oradell, supra. In the absence of any meaningful statutory standards, which might help secure the privacy interests at stake, I question whether we can assume, as a matter of law, that Government agents will be able to formulate for themselves constitutionally valid standards of review in examining, segregating, and cataloging the papers of the former President.
Nor does the possibility that, had Title I not been passed, appellant would perhaps use Government specialists to help classify and catalog his papers eliminate the objections to this intrusion. Had appellant, like all his recent predecessors, been permitted to deposit his papers in a Presidential library, Government archivists would have been working directly under appellant’s guidance and direction, not solely that of Congress or GSA. He, not Congress, would have established standards *536for preservation, to ensure that his privacy would be protected. Similarly, he would have been able to participate personally in the reviewing process and could thus assure that any governmental review of purely personal papers was minimized or entirely eliminated. He, not Congress, would have controlled the selection of which experts, if any, would have access to his papers. Finally, and most important, the “intrusion” would have been consented to, eliminating any constitutional question. But the possibility of a consent intrusion cannot, under our law, justify a nonconsensual invasion. Actual consent is required, cf. Schneckloth v. Bustamonte, 412 U. S. 218 (1973), not the mere possibility of consent under drastically different circumstances.
Finally, even if the Government agents are completely discreet, they are still Government officials charged with reviewing highly private papers and tape recordings. Unless we are to say that a police seizure and examination of private papers is justified by the “impeccable” record of a discreet police officer, I have considerable difficulty understanding how a compulsory review of admittedly private papers, in which there is no conceivable governmental interest, by Government agents is constitutionally permissible.
Ill
Bill of Attainder
A
Under Art. I, § 9, cl. 3, as construed and applied by this Court since the time of Mr. Chief Justice Marshall, Title I violates the Bill of Attainder Clause. In contrast to Title II of Pub. L. 93-526, the Public Documents Act, which establishes a National Study Commission to study questions concerning the preservation of records of all federal officials, Title I commands the Administrator to seize all tape recordings “involving] former President Richard M. Nixon” and all “Presidential historical materials of Richard M. Nixon . . . .” *537§§ 101 (a)(1), (b)(1). By contrast with Title II, which is general legislation, Title I is special legislation singling out one individual as the target.
Although the prohibition against bills of attainder has been addressed only infrequently by this Court, it is now settled beyond dispute that a bill of attainder, within the meaning of Art. I, is by no means the same as a bill of attainder at common law. The definition departed from the common-law concept very early in our history, in a most fundamental way. At common law, the bill was a death sentence imposed by legislative Act. Anything less than death was not a bill of attainder, but was, rather, “a bill of pains and penalties.” This restrictive definition was recognized tangentially in Marbury v. Madison, 1 Cranch 137, 179 (1803),29 but the Court soon thereafter rejected conclusively any notion that only a legislative death sentence or even incarceration imposed on named individuals fell within the prohibition. Mr. Chief Justice Marshall firmly settled the matter in 1810, holding that legislative punishment in the form of a deprivation of property was prohibited by the Bill of Attainder Clause:
“A bill of attainder may affect the life of an individual, or may confiscate his property, or may do both.” Fletcher v. Peck, 6 Cranch 87, 138. (Emphasis supplied.)
The same point was made 17 years later in Ogden v. Saunders, 12 Wheat. 213,286, where the Court stated:
“By classing bills of attainder, ex post facto laws, and laws impairing the obligation of contracts together, the *538general intent becomes very apparent; it is a general provision against arbitrary and tyrannical legislation over existing rights, whether of person or property.” (Emphasis supplied.)
More than 100 years ago this Court struck down statutes which had the effect of preventing defined categories of persons from practicing their professions. Cummings v. Missouri, 4 Wall. 277 (1867) (a priest); Ex parte Garland, 4 Wall. 333 (1867) (a lawyer). Those two cases established more broadly that "punishment” for purposes of bills of attainder is not limited to criminal sanctions; rather, “[t]he deprivation of any rights, civil or political, previously enjoyed, may be punishment . . . .” Cummings, supra, at 320.
Mr. Chief Justice Warren pointed out that the Constitution, in prohibiting bills of attainder, did not envision “a narrow, technical (and therefore soon to be outmoded) prohibition....” United States v. Brown, 381 U. S. 437, 442 (1965). To the contrary, the evil was a legislatively imposed deprivation of existing rights, including property rights, directed at named individuals. Mr. Justice Black, in United States v. Lovett, 328 U. S. 303, 315-316 (1946), stated:
“[The cases] stand for the proposition that legislative acts, no matter what their form, that apply either to named individuals or to easily ascertainable members of a group in such a way as to inflict punishment on them without a judicial trial are bills of attainder prohibited by the Constitution.” (Emphasis supplied.)
The only “punishment” in Lovett, in fact, was the deprivation of Lovett’s salary as a Government employee — an indirect punishment for his “bad” associations.
Under our cases, therefore, bills of attainder require two elements: first, a specific designation of persons or groups as subjects of the legislation, and, second, a Garland-Cummings-Lovett-Br own-type arbitrary deprivation, including depriva*539tion of property rights, without notice, trial, or other hearing.30 No one disputes that Title I suffers from the first infirmity, since it applies only to one former President. The issue that remains is whether there has been a legislatively mandated deprivation of an existing right.
B
Since George Washington’s Presidency, our constitutional tradition, without a single exception, has treated Presidential papers as the President’s personal property. This view has been congressionally and judicially ratified, both as to the ownership of Presidential papers, Folsom v. Marsh, 9 F. Cas. 342 (No. 4,901) (CC Mass. 1841) (Story, J., sitting as Circuit Justice), and, by the practice of Justices as to ownership of their judicial papers.
Congress itself has consistently legislated on this assumption. I have noted earlier that appropriation legislation has been enacted on various occasions providing for Congress’ purchase of Presidential papers. See Hearing before a Special Subcommittee of the House Committee on Government Operations on H. J. Res. 330, 84th Cong., 1st Sess., 28 (1955). Those hearings led Congress to establish a nonmandatory sys*540tem of Presidential libraries, again explicitly recognizing that Presidential papers were the personal property of the Chief Executive. In the floor debate on that measure, Congressman John Moss, a supporter of the legislation, stated: “Finally, it should be remembered that Presidential papers belong to the President . . . 101 Cong. Rec. 9935 (1955). Indeed, in 1955 in testimony pertaining to this proposed legislation, the Archivist of the United States confirmed:
“The papers of the Presidents have always been considered to be their personal property, both during their incumbency and afterward. This has the sanction of law and custom and has never been authoritatively challenged.” Hearing on H. J. Res. 330, supra, at 32.
Similarly, the GSA Administrator testified:
“As a matter of ordinary practice, the President has removed his papers from the White House at the end of his term. This has been in keeping with the tradition and the fact that the papers are the personal property of the retiring Presidents.” Id., at 14. (Emphasis supplied.)
In keeping with this background, it was not surprising that the Attorney General stated in an opinion in September 1974:
“To conclude that such materials are not the property of former President Nixon would be to reverse what has apparently been the almost unvaried understanding of all three branches of the Government since the beginning of the Republic, and to call into question the practices of our Presidents since the earliest times.” 43 Op. Atty. Gen. No. 1, pp. 1-2 (1974).
I see no escape, therefore, from the conclusion that, on the basis of more than 180 years’ history, the appellant has been deprived of a property right enjoyed by all other Presidents *541after leaving office, namely, the control of his Presidential papers.
Even more starkly, Title I deprives only one former President of the right vested by statute in other former Presidents by the 1955 Act — the right to have a Presidential library at a facility of his own choosing for the deposit of such Presidential papers as he unilaterally selects. Title I did not purport to repeal the Presidential Libraries Act; that statute remains in effect, available to present and future Presidents, and has already been availed of by former President Ford. The operative effect of Title I, therefore, is to exclude, by name, one former President and deprive him of what his predecessors — and his successor — have already been allowed. This invokes what Mr. Justice Black said in Lovett, could not be constitutionally done:
“Those who wrote our Constitution well knew the danger inherent in special legislative acts which take away the life, liberty, or property of particular named persons because the legislature thinks them guilty of conduct which deserves punishment. They intended to safeguard the people of this country from punishment without trial by duly constituted courts.” 328 U. S., at 317. (Emphasis supplied.)
But apart from Presidential papers generally, Title I on its face contemplates that even the former President’s purely family and personal papers and tape recordings are likewise to be taken into custody for whatever period of time is required for review. Some items, such as the originals of tape recordings of the former President’s conversations, will never be returned to him under the Act.
I need not, and do not, inquire into the motives of Congress in imposing this deprivation on only one named person. Our cases plainly hold that retribution and vindictiveness are not requisite elements of a bill of attainder. The Court *542appears to overlook that Mr. Chief Justice Warren in United States v. Brown, supra, concluded that retributive motives on the part of Congress were irrelevant to bill-of-attainder analysis. To the contrary, he said flatly: “It would be archaic to limit the definition of punishment to ‘retribution.’ ” Indeed, he expressly noted that bills of attainder had historically been enacted for regulatory or preventive purposes:
“Historical considerations by no means compel restriction of the bill of attainder ban to instances of retribution. A number of English bills of attainder were enacted for preventive purposes — that is, the legislature made a judgment, undoubtedly based largely on past acts and associations . . . that a given person or group was likely to cause trouble . . . and therefore inflicted deprivations upon that person or group in order to keep it from bringing about the feared event.” 381 U. S., at 458-459.
Under the long line of our decisions, therefore, the Court has the heavy burden of demonstrating that legislation which singles out one named individual for deprivation — without any procedural safeguards — of what had for nearly 200 years been treated by all three branches of Government as private property, can survive the prohibition of the Bill of Attainder Clause. In deciding this case, the Court provides the basis for a future Congress to enact yet another Title I, directed at some future former President, or a Member of the House or the Senate because the individual has incurred public disfavor and that of the Congress. Cf. Powell v. McCormack, 395 U. S. 486 (1969). As in United States v. Brown, Title I, in contrast to Title II, does “not set forth a generally applicable rule,” 381 U. S., at 450; it is beyond doubt special legislation doing precisely the evil against which the prohibitions of the “bills of attainder, ex post facto laws, and laws impairing the obligation of contracts . . .” were aimed. Ogden v. Saunders, 12 Wheat., at 286.
*543The concurring opinions make explicit what is implicit throughout the Court’s opinion, i. e., (a) that Title I would be unconstitutional under separation-of-powers principles if it applied to any other President; (b) that the Court’s holding rests on appellant’s being a “legitimate class of one,” ante, at 472; and (c) that the Court’s holding “will not be a precedent.” Ante, at 486.
Nothing in our cases supports the analysis of Mr. Justice Stevens, ibid. Under his view, appellant’s resignation and subsequent acceptance of a pardon set him apart as a “ 'legitimate class of one.’ ” The two events upon which he relies, however, are beside the point. Correct analysis under the Bill of Attainder Clause focuses solely upon the nature of the measure adopted by Congress, not upon the actions of the target of the legislation. Even if this approach were analytically sound, the two events singled out are relevant only to two possible theories: first, that appellant is culpably deserving of punishment by virtue of his resignation and pardon; or second, that appellant’s actions were so unique as to justify legislation confiscating his Presidential materials but not those of any other President. The first point can be disposed of quickly, since the Bill of Attainder Clause was, of course, intended to prevent legislatively imposed deprivations of rights upon persons whom the Legislature thought to be culpably deserving of punishment.
The remaining question, then, is whether appellant’s “uniqueness” permits individualized legislation of the sort passed here. It does not. The point is not that Congress is powerless to act as to exigencies arising during or in the immediate aftermath of a particular administration; rather, the point is that Congress cannot punish a particular individual on account of his “uniqueness.” If Congress had declared forfeited appellant’s retirement pay to which he otherwise would be entitled, instead of confiscating his Presidential materials, it would not avoid the bill-of-attainder prohibition to say that appellant was guilty of unprecedented actions *544setting him apart from his predecessors in office. In short, appellant’s uniqueness does not justify serious deprivations of existing rights, including the statutory right abrogated by Title I to establish a Presidential library.
The novel arguments advanced in the several concurring opinions serve to emphasize how clearly Title I violates the Bill of Attainder Clause; Mr. Justice Stevens although finding no violation of the Clause, admirably states the case which, for me, demonstrates the unconstitutionality of Title I:
“The statute before the Court does not apply to all Presidents or former Presidents. It singles out one, by name, for special treatment. Unlike all former Presidents in our history, he is denied custody of his own Presidential papers; he is subjected to the burden of prolonged litigation over the administration of the statute; and his most private papers and conversations are to be scrutinized by Government archivists. The statute implicitly condemns him as an unreliable custodian of his papers. Legislation which subjects a named individual to this humiliating treatment must raise serious questions under the Bill of Attainder Clause.” Ante, at 484.
IV
The immediate consequences of the Court’s holding may be discounted by some on the ground it is justified by the uniqueness of the circumstances — in short, that the end justifies the means — and that, after all, the Court’s holding is really not to be regarded as precedent. Yet the reported decisions of this Court reflect other instances in which unique situations confronted the Judicial Branch — for example, the alleged treason of one of the Founding Fathers. United States v. Burr, 25 F. Cas. 187 (No. 14,694) (CC Va. 1807). Burr may or may not have been blameless; Father Cummings and Lawyer Garland, in common with hundreds of thousands of others, may have been technically guilty of “carrying on *545rebellion” against the United States. But this Court did not weigh the culpability of Cummings, Garland, or of Lovett or Brown in according to each of them the full measure of the protection guaranteed by the literal language of the Constitution. For nearly 200 years this Court has not viewed either a “class” or a “class of one” as “legitimate” under the Bill of Attainder Clause.
It may be, as three Justices intimate in their concurring opinions, that today’s holding will be confined to this particular “class of one”; if so, it may not do great harm to our constitutional jurisprudence but neither will it enhance the Court’s credit in terms of adherence to stare decisis. Only with future analysis, in perspective, and free from the “hydraulic pressure” Holmes spoke of, will we be able to render judgment on whether the Court has today enforced the Constitution or eroded it.
Later, I will discuss the importance of the legislation’s applicability to only one ex-President.
In this, Mr. Justice Miller was but expressing the earlier opinion of Madison, who declared in The Federalist No. 48, p. 334 (J. Cooke ed., 1961):
“The legislative department derives a superiority in our governments from other circumstances. Its constitutional powers being at once more extensive and less susceptible of precise limits, it can with the greater facility, mask under complicated and indirect measures, the encroachments which it makes on the co-ordinate departments.”
Statutes relating to departments or agencies created by Congress frequently are phrased in mandatory terms. For example, in the 1949 legislation creating the General Services Administration, Congress provided as follows:
“The Administrator is authorized and directed to coordinate and provide for the . . . efficient purchase, lease and maintenance of . . . equipment by Federal agencies.” 40 U. S. C. § 759 (a).
Even with respect to international relations, Congress has affirmatively imposed certain requirements on the Secretary of State:
“The Secretary of State shall furnish to the Public Printer a correct copy of every treaty between the United States and any foreign government _” 22 U. S. C. § 2660.
Cf. Mr. Justice White’s discussion in United States v. Brewster, 408 U. S. 501, 558 (1972) (dissenting opinion), where he spoke of the *509“evil” of “executive control of legislative behavior (Emphasis supplied.)
This Presidential prerogative has not been limited to foreign affairs, where, of course, secrecy and confidentiality may be of the utmost importance. See A. Bickel, The Morality of Consent 79 (1975); W. Taft, The Presidency 110 (1916).
The Court’s references to the historical understanding of separation-of-powers principles omit a crucial part of that history. Madison’s statements in The Federalist No. 47 as to one department’s exercising the “whole power” of another department do not purport to be his total treatment of the subject. The Federalist No. 48, two days later, states the central theme of Madison’s view:
“It is equally evident, that neither [department] ought to possess directly or indirectly, an overruling influence over the others in the administration of their respective powers.” The Federalist No. 48, p. 332 (J. Cooke ed. 1961). (Emphasis supplied.)
Indeed, Madison expressly warned at length in No. 48 of the inevitable *512dangers of “encroachments” by the Legislative Branch upon the coordinate departments o.f Government.
But aside from the Court’s highly selective discussion of the Framers’ understanding, the Court cannot obscure the fact that this Court has never required, in order to show a separation-of-powers violation, that Congress usurped the whole of executive power. Any such requirement was rejected by the Court in Buckley v. Valeo, 424 U. S. 1 (1976). There, we held that Congress could not constitutionally exercise the President’s appointing powers, even though under that statute the President had the power to appoint one-fourth of the Federal Election Commission members, and even though the President had “approved” the statute when he signed the bill into law.
Nowhere is the standard clarified in the majority’s opinion. We are left to guess whether only a “potential for disruption” is required or whether “undue disruption,” whatever that may be, is required.
The federal parties filed three briefs in Buckley. The main brief, styled the “Brief for the Attorney General as Appellee and for the United States as Amicus Curiae,” explicitly stated that the method of appointment of four of the members of the Commission was unconstitutional. See pp. 6-7, 110-120. The Attorney General signed this portion of the brief as a party (see pp. 2, 103 n. 65). The Executive Branch therefore made it clear that, in its view, the statute was unconstitutional to the extent it reposed appointing powers in Congress. The second brief, styled the “Brief for *513the Attorney General and the Federal Election Commission/’ generally defended the Act but took no position concerning the method of appointing the Commission. See p. 1 n. 1. The third brief was filed by the Commission on its own behalf only; it defended the appointment procedures, but it was not joined by the Attorney General and did not express the view of the President or of any other portion of the Executive Branch.
As to congressional papers, see supra, at 510-511. Despite the Constitution’s silence as to the papers of the Legislative Branch, this Court had no difficulty holding those papers to be protected from control by other branches. See also Mr. Justice Brennan’s dissenting opinion in United States v. Brewster, 408 U. S. 501, 532-533 (1972), where he quotes approvingly from Kilbourn v. Thompson, 103 U. S. 168 (1881), and Coffin v. Coffin, 4 Mass. 1 (1808). In both of those cases, written materials by legislators were deemed to be protected by legislative immunity from intrusion or seizure.
This discretion was exercised, as we have seen, by President Washington in the face of a congressional demand for production of his work-papers.
Obviously, official documents fall into an entirely different category and are not involved in this case.
Appellees, of course, would view that sort of intrusion as an intra-branch confrontation, since United States Marshals are officials of the Executive Branch, at least so long as the District Judge simply ordered the Marshals to take custody of and to review the documents without turning them over to the court. This is, of course, sheer sophistry.
Of course, United States v. Nixon pertained only to the setting of Judicial-Executive conflict. Nothing in our holding suggests that, even if Congress needed Presidential documents in connection with its legislative functions, the constitutional tradition of Presidential control over Presidential documents in the face of legislative demands could be abrogated. We expressly stated in Nixon that “[w]e are not here concerned with the balance between . . . the confidentiality interest and congressional demands for information . . . .” 418 U. S., at 712 n. 19.
In his concurring opinion, Mr. Justice Powell concludes that Title I was addressed essentially to an “emergency” situation in the wake of appellant’s resignation. But his opinion does not present any analysis as to whether this particular legislation, not some other legislation, is necessary to achieve that end. Since Title I commands confiscation of *517all materials of an entire Presidential administration, Title I was simply not drafted to meet the specific emergency it purports to address. Besides omitting any discussion justifying the need for Title I, Mr. Justice Powell’s opinion relies entirely on the possibly limiting regulations to be promulgated at some future point by the GSA Administrator, which will protect “all constitutional and legal rights . . . .” Ante, at 497. This conclusion, of course, begs the precise question before us, which is whether the act of congressionally mandated seizure of all Presidential materials of one President violates the Constitution.
Civil service statutes aside, we know now that an executive official cannot replace all of his underlings on the basis of a patronage system. Thus, as a matter of constitutional law, a Chief Executive would not be at liberty to replace all Executive Branch officials with persons who, for political reasons, enjoy the President’s trust and confidence. Elrod v. Burns, 427 U. S. 347 (1976).
1 cannot accept the argument pressed by appellees that review is rendered harmless by the fact that many of the documents may not be protected by Presidential privilege. How “harmless” review justifies manifestly “harmful” review escapes me.
President Truman, for one, objected to Congress’ efforts to coerce him after he was no longer in office in connection with matters pertaining to his administration. See infra, at 522.
It would be the height of impertinence, after all, to serve a legislative subpoena on an outgoing President as he is departing from the inauguration of his successor. So too, the people would rightly be offended, and more important, so would the Constitution, by a congressional resolution, designed to ensure the smooth functioning of the Executive Branch, requiring a former President, upon leaving office, to remain in *520Washington, D. C., in order to be available for consultations with his successor for a prescribed period of time.
The fact that the President signs a bill into law, and thereafter defends it, without more, does not mean, of course, that the policy embodied in the legislation is that of the President, nor does it even mean that the President personally approves of the measure. When signing a bill into law, numerous Presidents have actually expressed disagreement with the legislation but felt constrained for a variety of reasons to permit the bill to become law. President Franklin D. Roosevelt repudiated the “Lovett Rider” later struck down by this Court in United States v. Lovett, 328 U. S. 303, 325 (1946) (Frankfurter, J., concurring). President Ford did not request this legislation in order to assure the effective functioning of the Executive Branch.
Since by definition the concern is with former Presidents, I see no distinction in Congress’ seeking to compel the appearance and testimony *523of a former President and in, alternatively, seeking to compel the production of Presidential papers over the former President’s objection.
Clients asserting the attorney-client privilege have not, up to now, been foreclosed from interposing the privilege unless a similarly situated client is willing to support the particular claim.
The District Court concluded its discussion of the privacy challenge as follows: “We would be less than candid were we to state that we find it as easy to dispose of Mr. Nixon’s privacy claims as his claim of presidential privilege.” 408 F. Supp., at 367.
Although the District Court expressly concluded that the former President had a “legitimate expectation” that his Presidential materials would not be subject to “comprehensive review by government personnel without his consent,” id., at 361, the Court nonetheless deemed the compulsory intrusion permissible given the constitutionality of the federal wiretap statute, 18 U. S. C. §§ 2510-2520, which of course permits substantial governmental intrusions into the privacy of individuals. Not only is this analogy imperfect, as the District Court itself admitted, 408 F. Supp., at 364, but this analysis fails to apply the “exacting scrutiny” called for by our decisions. Above all, the present statute fails to provide any of the *527stringent safeguards, including a warrant, mandated by Title III of the Omnibus Crime Control and Safe Streets Act of 1968. Indeed, the District Court flatly admitted as much. Ibid.
Administrative efficiency is obviously a highly desirable goal. See, e. g., Dixon v. Love, 431 U. S. 105, 114 (1977); Mathews v. Eldridge, 424 U. S. 319, 347-349 (1976). However, I am constrained to recall that “administrative efficiency” has not uniformly been regarded as of “overriding importance.” Indeed, claims of administrative efficiency have been swiftly dismissed at times as mere “bald assertion [s].” Richardson v. Wright, 405 U. S. 208, 223 (1972) (BrenNAN, J., dissenting). Numerous other opinions have held that individual interests, including the right to welfare payments, “clearly outweigh” government interests in promoting “administrative efficiency,” Goldberg v. Kelly, 397 U. S. 254 (1970) (opinion of BreNNAN, J.). And, Mr. Justice Marshall in Shapiro v. Thompson, 394 U. S. 618, 634 (1969), stated that when “fundamental” rights are at stake, such as the “right to travel,” government must demonstrate a “compelling” interest, not merely a “rational relationship between [the underlying statute] and [the] . . . admittedly permissible state objectives . . . .”
The initial interest in preserving the materials for judicial purposes has diminished substantially. Since the Special Prosecutor has disclaimed any further interest in the materials for purposes of possible criminal investigations, the only conceivably remaining judicial need is to preserve the materials for possible use in civil litigation between private parties. The admittedly important interests in the enforcement of the criminal law, recognized in United States v. Nixon, are no longer pressed by the Government.
If there were a particularized need, the statute suffers from greater overbreadth than others we have invalidated.
At the time Title I was passed, appellant had made tentative arrangements with the University of Southern California in Los Angeles for the establishment of a Presidential library, under the terms of the Presidential Libraries Act. App. 167-168. That has now ripened into a formal agreement so that in the event Title I is invalidated, appellant’s historical materials will be housed in a facility on the USC campus under terms applicable to other Presidential libraries of past Presidents.
The Court’s refusal to afford constitutional protection to such commercial matters as bank records, California, Bankers Assn. v. Shultz, 416 U. S. 21 (1974), or drug prescription records, Whalen v. Roe, 429 U. S. 589 (1977), only serves to emphasize the importance of truly private papers or communications, such as a personal diary or family correspondence. These private papers lie at the core of First and Fourth Amendment interests.
The fact that GSA initially secured possession of the Presidential papers through the agreement with the former President does not change the fact that the agency was commanded by Congress to take exclusive custody of and retain all Presidential historical materials. Moreover, everyone admits that the Act contemplates a careful screening process by Government agents. The fact that the governmental intrusion is noncriminal in nature does not, of course, render the Fourth Amendment’s prohibitions inapplicable. See South Dakota v. Opperman, 428 U. S. 364 (1976).
“The constitution declares that 'no bill of attainder ox ex post facto law shall be passed.’
“If, however, such a bill should be passed and a person should be prosecuted under it; must the court condemn to death those victims whom the constitution endeavors to preserve?” Marbury v. Madison, 1 Cranch, at 179.
Title I fails to provide any procedural due process safeguards, either before or after seizure of the Presidential materials. There is no provision whatever permitting appellant to be heard in the decisionmaking process by which GSA employees will determine, with no statutory standards to guide them, whether particular materials have “general historical value.” No time restraints are placed upon GSA’s decisionmaking process, even though this Court has consistently recognized that, when dealing with First Amendment interests, the timing of governmental decisionmaking is crucial. E. g, Freedman v. Maryland, 380 U. S. 51 (1965); Marcus v. Search Warrant, 367 U. S. 717 (1961,) Under those holdings, any statute which separates an individual, against his will, from First Amendment protected materials must be strictly limited within a time frame. Title I, in contrast, places no limits with respect to GSA’s retention of custody over appellant’s papers; three years have already elapsed since seizure of the papers in question.