This post has been co-written by the following, all of whom have served in the Department of Justice’s Office of Legal Counsel: David Barron (Harvard Law School), Walter Dellinger (Duke Law School), Dawn Johnsen (Indiana School of Law, Bloomington), Neil Kinkopf (Georgia State Law School), Marty Lederman (Georgetown Universty Law Center), Chris Schroeder (Duke Law School), Richard Shiffrin (University of the Pacific McGeorge School of Law), Michael Small (Akin, Gump, Strauus, Hauer & Feld, LLP).

President Bush has made extensive use of signing statements to announce his constitutional doubts about numerous statutory enactments and to signal his intent either to refuse to enforce such laws or to construe and implement them in a manner consistent with his constitutional views. This practice first gained notoriety with the President’s signing statement regarding the McCain Amendment back in December, and has received much more attention thanks largely to the valuable archival work of Phillip Cooper, Christopher Kelley and Charlie Savage.

The American Bar Association recently joined the debate with the release of a Report by a bipartisan task force on presidential signing statements and the separation of powers. As one of us has noted previously, a number of aspects of the Report are very timely and worthwhile. In particular, we agree with the Report’s emphasis on the need for transparency whenever the executive branch declines to execute a statute or construes that statute in a manner that appears to be contrary to its text and congressional intent. But we were surprised by, and disagree with, certain of the Report's central conclusions, namely:

(1) that "the ‘take care’ obligation of the President requires him to faithfully execute all laws” (page 19), including, apparently, those that are unconstitutional; and

(2) that if a President believes any provision of a bill is unconstitutional, he is obliged either (a) to veto the entire bill, or (b) to sign the bill and enforce the unconstitutional provisions.

As we explain below, the first conclusion is untenable. There is a long history of Presidents concluding that, in certain limited circumstances, it is more consistent with their constitutional obligations to refuse to enforce an unconstitutional law than to enforce it. As just the most obvious example, some laws might be plainly unconstitutional under governing and uncontroverted Supreme Court precedent. We do not believe there is a serious case to be made for a constitutional obligation to enforce all such provisions.

But the main target of the ABA Report appears to be not so much these easy cases, but instead cases such as the recent McCain Amendment categorically prohibiting the cruel treatment of prisoners in U.S. custody. President Bush signed the omnibus appropriations bill of which this provision was one small part and at the same time issued a signing statement obliquely suggesting that he may reserve the power to make exceptions pursuant to his authority as Commander-in-Chief. There are serious problems with the views expressed in such Bush signing statements, but the Report misses the mark in identifying them. Like most misdiagnoses, the Report may have the doubly negative effect of concentrating attention on a phony problem—the issuance of signing statements that long have been used to signal the President’s belief that some aspect of a piece of legislation is unconstitutional—while at the same time deflecting attention from a very real problem, namely, the unjustifiable arrogation of power that President Bush has asserted and continues to assert in office.

Below we identify three common objections to the use of signing statements (not all of them directly attributable to the ABA Report) and explain why they are misplaced. Then we discuss four real problems reflected in the content of presidential signing statements and accompanying administrative practices in this Administration.

First, however, a few words to try to unpack precisely what the ABA Report finds objectionable:

Although the Report concludes (p.1) that it is "contrary to the rule of law and our constitutional system of separation of powers" to issue signing statements claiming an authority, or stating an intention, to disregard or decline to enforce part of a law that the President has signed, the full Report seems ambiguous in its statement of the problem. Most of the Report reads as though the Task Force is not so much opposed to signing statements, as such, but that it is instead aiming its sights at the non-enforcement of laws based on constitutional objections. If so, then the signing statements would merely be a signal of the intent to refuse to enforce, and not themselves the problem.

On the other hand, at page 27, the Report oddly suggests that the only problem it is addressing is related to signing statements themselves (and the failure to veto bills with constitutionally dubious provisions), and not to nonenforcement based on constitutional objections more broadly. The Report states that the Task Force is not addressing “how the President should respond if Congress overrides a veto motivated by his constitutional concerns,” or even “what should be done if the President, in the absence of a signing statement, nevertheless fails to enforce a law,” which would presumably include failing to enforce a provision signed by a prior President.

In other words, the Task Force appears, ironically enough, to be agnostic on the central constitutional problem in this Administration: the President’s assertion of a power to ignore statutes that have been on the books for years, that were widely acknowledged to be constitutional when they were enacted (including by the President when he signed the law), that for many years thereafter have been implemented by the executive branch, and that have never been called into question by the Supreme Court. Regarding situations where the Bush Administration (secretly) asserts the right to ignore existing basic framework statutes, the ABA Task Force is strangely silent. Counterintuitively, it aims its sights only at those occasions when the President actually signals his intent to refuse to enforce new statutes.

By thus calling into question the inclusion of some words in a document, rather than the substance and intentions of the views about executive power that President Bush’s words express, the Task Force’s focus is backward. As explained below, to the extent a President is refusing to enforce a statute, it is much better that he announce his attention to do so publicly, rather than doing so secretly.)

The Report thus may draw attention away from the truly serious problems that the Bush signing statements do reflect. The Report discusses some of those problems (e.g., the absence of transparency), but it studiously avoids addressing the most significant problem with the Bush signing statements, as well as with other failures to enforce preexisting statutes—namely, that the substance of many of this President's constitutional objections is wrong and threatens to dangerously expand the powers of the President in a manner that fails to respect the checks and balances of our constitutional system.

In this regard, the Report may simply be mirroring the larger public discussion of the practice of issuing signing statements, a discussion that has tended to fixate on the phenomenon of the statements themselves. Three commonly heard objections recur in that discussion. Each of them misses the mark:

First, and most importantly, some objections to recent signing statements appear to be premised on the notion that the President is categorically prohibited from refusing to enforce a statute that he determines to be unconstitutional. But such a categorical prohibition is belied by a long history of Presidents declining, in certain limited circumstances, to enforce statutes that they deem unconstitutional. As the Task Force Report notes, constitutional nonenforcement has been occurring since at least the Grant Administration. Indeed, the practice began no later than the Buchanan Administration, if not earlier still—in 1801, President Jefferson ordered the cessation of prosecutions under the Sedition Act, a statute that he viewed as unconstitutional (a judgment that the Supreme Court would not confirm until 163 years later). And in 1983, the Supreme Court acknowledged with apparent approval in INS v. Chadha the “not uncommon” practice of Presidents approving legislation “containing parts which are objectionable on constitutional grounds.”

To be sure, the ABA Report is correct that the practice has increased in modern times, and some of that increase can be attributed to presidential abuse. But much of the increase is a function of Congress’s increased use of omnibus legislation that includes, among literally hundreds of constitutionally unobjectionable provisions, a handful of provisions that might be unconstitutional, including some that are clearly invalid under governing Supreme Court precedent. In such cases, at the very least, a refusal to enforce has been and should continue to be uncontroversial.

We agree wholeheartedly that a President cannot simply choose not to enforce whichever statutes he does not like. A President may not exercise a dispensing power—in effect a type of “line-item-veto”—to ignore statutes that he thinks are unwise, or wrong, or politically inexpedient. See United States ex rel. Stokes v. Kendall. The Task Force thus is absolutely right (p.23) when it quotes former OLC Assistant Attorney General Rehnquist to the effect that presidential "impoundment" of appropriated funds is unconstitutional, and that the power to execute laws does not give the President the authority to decline to execute them. The President has an obligation under Article II to faithfully execute the laws.

But that "Take Care" obligation includes a responsibility, above all, to faithfully execute the Constitution.Thus, in some limited and relatively rare circumstances, the President might best fulfill his constitutional obligation by refusing to enforce a statutory provision that he considers to be unconstitutional.

This decidedly does not mean that the President can or should always choose noncompliance every time his personal view is that a particular provision is unconstitutional. The question of when the President ought to refuse to enforce a statute—rather than the more common practice of enforcing the statute but then not defending it in court—is a very difficult and sensitive one, and has inspired a rich literature. See, for example, this colloquy a few years back in Law & Contemporary Problems in which two of us (Johnsen and Barron) participated.

The issue was succinctly addressed by the Office of Legal Counsel in the Clinton Administration, in Assistant Attorney General Walter Dellinger’s 1994 memorandum to Abner Mikva. (Disclosure: We each worked at OLC under Clinton and several of us worked on the 1994 memorandum.):

Th[e] decision [to refuse enforcement on constitutional grounds] is necessarily specific to context, and it should be reached after careful weighing of the effect of compliance with the provision on the constitutional rights of affected individuals and on the executive branch's constitutional authority. Also relevant is the likelihood that compliance or non-compliance will permit judicial resolution of the issue. That is, the President may base his decision to comply (or decline to comply) in part on a desire to afford the Supreme Court an opportunity to review the constitutional judgment of the legislative branch.

We think this basically gets the question right, at least as a general matter, and subject to certain refinements and elaborations contained in the Johnsen and Barron articles linked above.

Accordingly, we are on common ground with the ABA Task Force in several important respects, namely:

? Nonenforcement should never be the President's first resort when met with a constitutionally dubious statute—and should rarely be the last resort, either. The Bush Administration’s seemingly cavalier assertions of the authority to refuse to enforce laws as Congress has written them places it at odds with its predecessors of both political parties.

? The President is obliged to work diligently with Congress to eliminate any constitutional doubts during the legislative process. As Walter Dellinger wrote in that 1994 memo: "When bills are under consideration by Congress, the executive branch should promptly identify unconstitutional provisions and communicate its concerns to Congress so that the provisions can be corrected. Although this may seem elementary, in practice there have been occasions in which the President has been presented with enrolled bills containing constitutional flaws that should have been corrected in the legislative process."

? The President, like a Court, should treat Congress's contrary constitutional judgment seriously, and begin analysis of a statute with a presumption of constitutionality. As we have written in a related context, Executive branch legal analysis “ should be thorough and forthright, and it should reflect all legal constraints, including the constitutional authorities of the coordinate branches of the federal government—the courts and Congress—and constitutional limits on the exercise of governmental power."

? The Supreme Court has a special role in establishing constitutional meaning in our system, and therefore the President should rarely refuse to enforce a law without some confidence that doing so would not be inconsistent with the Court’s own views.

? And, relatedly, as Dawn Johnsen's article explains, the President should typically act in a way that promotes judicial resolution of the constitutional dispute between the political branches. The option of enforcement and then a refusal to defend should always be carefully considered as a default rule, even if (as David Barron’s article explains) that fallback might not always be the optimal solution.

Briefly summarized, we think nonenforcement on any seriously contested question of constitutional law should be the rare exception, a rule of thumb that coincides with Executive practice prior to this Administration.

In contrast, the view that nonenforcement on constitutional grounds should be condemned under any circumstances strikes us as not only historically anomalous, but untenable, particularly in light of the ever-growing number of clearly unconstitutional statutory provisions duly enacted by Congress and sitting on the books. Should Jefferson really have permitted sedition prosecutions to continue? Should the Executive branch enforce criminal laws, still on the books, that make it unlawful to distribute information about abortions? Enforce social-security and welfare laws that discriminate on the basis of sex in a manner patently inconsistent with Supreme Court case law? Implement one-house vetoes, even after Chadha? How would such enforcement possibly constitute faithful execution of the Constitution?

Second, despite all the negative attention that signing statements have received in the past months, there is nothing inherently wrong with signing statements as such—including those that contain constitutional objections. Thus, for example, the statement of the ABA President when he unveiled the Report last week—that “the threat to our Republic posed by presidential signing statements is both imminent and real unless immediate corrective action is taken”—misses the point. So does section 5 of Senator Specter's new bill, which would create a cause of action for the Senate or the House to sue for a declaratory judgment on "the legality of any presidential signing statement."

Wholly apart from the serious standing and case-or-controversy problems this proposal would raise, it is substantively misguided: There is absolutely nothing unlawful about any signing statement. The constitutional problem arises when the President executes, or fails to execute, a statute.

Presidents have used such statements throughout our history. As Walter Dellinger explained in another OLC memo:

[S]uch statements may on appropriate occasions perform useful and legally significant functions. These functions include (1) explaining to the public, and particularly to constituencies interested in the bill, what the President believes to be the likely effects of its adoption, (2) directing subordinate officers within the Executive Branch how to interpret or administer the enactment, and (3) informing Congress and the public that the Executive believes that a particular provision would be unconstitutional in certain of its applications, or that it is unconstitutional on its face, and that the provision will not be given effect by the Executive Branch to the extent that such enforcement would create an unconstitutional condition.

Dellinger’s third category—a statement announcing the intent not to enforce an unconstitutional provision—obviously is the most controversial. But again, it’s hardly unprecedented.

Indeed, the ABA Report gets it exactly backwards. The signing statement is a good thing: a manifestation of the Executive’s intentions that helps us to understand the heart of the problem. If the President has decided to decline to enforce a statute because it’s unconstitutional (see above) then it is much better that he tell the Congress and the public of his intentions, rather than keep it secret, because in that case the checks and balances of the constitutional system can be set to work.

A close reading of the ABA Report suggests, however, that the Task Force’s real concern is not so much the signing statement, as such, but a purported constitutional failure at the moment of presentment. The Task Force view appears to be that if a President is presented with a bill that he knows contains an unconstitutional provision he has a binary choice: veto it or sign it and enforce it fully. (See the Task Force's third recommended, "either/or" Resolution.)

There is an intuitive appeal to this notion. After all, if the President determines that a bill presented to him contains an unconstitutional provision, doesn’t he have an obligation under the Take Care Clause to veto the bill, rather than to sign it and then refuse to enforce the provision?

We believe there is such an obligation if the entire bill is facially unconstitutional. But that doesn't describe the recurring problem in modern government. Much more often, the constitutionally objectionable provisions are included in important omnibus bills, such as an appropriations bill, containing numerous other valuable or essential provisions. And it is with respect to these omnibus bills that Presidents generally have resorted to the signing statement and nonenforcement, rather than using the veto.

One example: President Roosevelt signed the Lend Lease Act, despite his conclusion that one of its provisions was unconstitutional, because he believed the Act vital to the success of World War II. The provision he believed unconstitutional authorized Congress to rescind, by concurrent resolution, specified authorities granted to the President. President Roosevelt’s constitutional interpretation was vindicated over 40 years later in INS v. Chadha. Congress continues to this day repeatedly to enact provisions that seek to expand congressional power in violation of Chadha.

One might argue that this longstanding practice should be nipped in the bud—that Congress should be deterred from including such constitutionally dubious provisions in important omnibus bills and that a series of vetoes is just the way to do it. Perhaps that would be a salutary development, but we do not think the Constitution compels such a result.

The Task Force suggests that the Presentment Clause prohibits the President from using such signing statements. But that concern is off the mark. When the President signs a bill presented to him it becomes positive law—all of it, even the constitutionally objectionable provisions—and thus the Presentment Clause is satisfied. The fact that the President asserts a right not to enforce it does not mean that it is wiped off the books. There is, in other words, no “line-item veto.” An example should prove the point: If President Bush had had the power to “line-item veto” the McCain Amendment, it would never have become law, and would never have bound federal interrogation practices. But because he did not veto it, it is an actual statute: It binds the conduct of executive branch actors in the absence of a presidential directive not to enforce it, and it can and will be enforced by future Presidents who disagree with President Bush’s view of the Commander-in-Chief Clause (or if the Supreme Court were to declare that it is constitutional).

Moreover, even if one thinks that it would be good policy for the President to veto all bills containing unconstitutional provisions, it will never happen. Presidents will not begin to veto finely wrought and hard-fought legislation of any importance just because two or three provisions out of a thousand contain a Chadha violation, or unduly impinge on Executive authority. And if it ever came to pass that Presidents did view the veto as their only option, they often would not exercise it. If they took the Task Force Report’s either/or resolution to heart, they often would swallow their constitutional objections and sign the bill without objection, after which the only options would be unannounced nonenforcement of the dubious provisions or enforcement of unconstitutional laws. That result would, in our judgment, be perverse, depriving the public and the Congress of useful information for no good purpose.

There's yet another problem with this sign-it-or-veto-it view: In many such cases, the President's view, reflected in signing statements, is not that entire statutory provisions are facially unconstitutional, but merely that the laws might be unconstitutional in some future hypothetical applications. Does even that possibility of some future constitutional concern require a veto? And if not, would the President later be required to enforce the law in an unconstitutional manner?

In sum, we think longstanding practice represents the better view. As Walter Dellinger put the point in an Op-Ed in today’s New York Times:

When a bill with a thousand provisions includes one that is unconstitutional, the Constitution does not force the president to choose between two starkly unpalatable options: veto the entire bill or enforce an unconstitutional provision. A signing statement that announces the president’s intention to disregard the invalid provision offers a valuable, and lawful, alternative.

(As explained above, however, we do believe that before resorting to non-enforcement, the President should explore alternatives less threatening to the lawmaking process, including corrections to legislation before passage, interpreting genuinely ambiguous provisions to avoid constitutional problems, and consideration of the possibility of a veto and reenactment without the unconstitutional provision.)

Third, the Reagan Administration promoted a practice of using signing statements to create legislative (or “executive”) history, in the hopes that courts would accord weight to such executive interpretations in ascertaining the meaning of statutes. This was the initiative discussed in the Alito hearings, because now-Justice Alito had written a memo on this topic when he worked at OLC. This is a primary focus of Senator Specter's bill, which would purport to prohibit state and federal courts from relying on or deferring to a presidential signing statement as a source of authority in determining the meaning of any Act of Congress.

We don’t think this issue will amount to much, because we doubt that courts will any time soon come to treat presidential intent as being as significant as congressional intent in construing statutes. (For more on this question, see the 1993 Dellinger memo.) In any event, it should be a non-issue in the current debate, because the Bush Administration statements do not appear designed to influence judicial interpretation: Indeed, the Bush Administration does not appear to be at all interested in having courts consider the constitutionality of the statutes that they are refusing to enforce.

* * * *

So, if those are not the problems with the current Administration’s practices, what are the problems? We already have discussed some of our concerns above, but to summarize here, we see at least four:

First, in this Administration nonenforcement appears to be a strategy of first resort, not last. The guidelines that OLC set forth during the Clinton Administration (discussed above) place a focus on the Constitution’s structure. Those guidelines make clear that the President is to act in ways that respect the important roles of Congress and the courts in the process of constitutional interpretation and the resolution of constitutional controversy. The frequent and cavalier declarations of constitutional objections by the Bush Administration demonstrate that it pays little or no heed to these fundamental constitutional values.

Second, in many cases this Administration’s signing statements do not plausibly reflect legislative intent.

More often than not, the Administration has justified such constructions as necessary in order to avoid a serious constitutional question. But even if the avoidance canon applies to the President’s interpretations just as strongly as it does to the judiciary’s interpretations, it is triggered only where the statute in question is genuinely ambiguous. See, e.g., United States v. Oakland Cannabis Buyers’ Coop. This Administration has too frequently misused the avoidance canon to distort the meaning of statutory provisions that were not ambiguous—both with respect to provisions President Bush signed into law and provisions long on the books. (Concededly, previous Administrations have also engaged in dubious use of the canon, albeit not with the frequency or audacity of the current Administration. However, in many (but not all) such cases, private individuals were not affected, and Congress knew exactly what was going on and had mechanisms to counter the nonenforcement.)

If the President believes that a statute, fairly construed, would be unconstitutional, he should simply say so, and consider whether to enforce it or to take some other action to address his constitutional concern, rather than couching his objection as an implausible form of statutory “construction.”

Third, the whole point of such signing statements—the reason that making them is actually a valuable practice (see above)—is that they make transparent the President’s intent to decline to enforce statutes in the manner contemplated by Congress. But President Bush’s statements for the most part do not serve this function. Many of the objections are written in such general and opaque terms, and with resort to vague assertions about an intent to “construe” the provisions in conformity with the Commander-in-Chief Clause, the “unitary executive,” etc., that it is impossible to know just what they mean in terms of how the Administration is implementing the statutes in question. According to Prof. Cooper, in President Bush's first term alone he offered 505 constitutional objections to various statutory provisions, and many of those objections applied to multiple provisions within a particular bill. This might mean that the Executive is refusing to implement hundreds of statutes enacted since 2001 (and many enacted prior to that date, too), or construing them in an implausible and unexpected way—or, then again, it might not. Congress and the public are offered no clear understanding of the legal theory of unconstitutionality, or of precisely which statutory provisions will not be enforced, under what circumstances, and why. The statements are, instead, mere placeholders, with respect to a vast number of statutory provisions, signaling that the Administration reserves the right not to enforce numerous unspecified provisions.

Moreover, the President is not telling Congress when he does refuse to enforce (or when he construes the statute in a manner that the legislature could not possibly have contemplated). A President may seek to enforce his own conception of the Constitution, even if it is a sharp break with the past. But when he does so, he is constitutionally obliged to do so in broad daylight, with adequate opportunity for the other branches and the public to understand the legal theory and the practice and to respond accordingly. Checks and balances can't possibly work if the revolution is occurring in secret, or if the Administration publicly insists that all is business as usual, that all statutes and treaties are being implemented as they always have been, while simultaneously "implementing" such statutes in a manner that comes as a great surprise to Congress and the public. Precisely in order to deal with these problems—to provide Congress the opportunity to evaluate, oversee, and check the President's nonenforcement—federal law already requires that such decisions be disclosed to Congress. Section 530D(a)(1)(A)(i) of title 28 of the U.S. Code provides that the Attorney General "shall submit to the Congress a report of any instance in which the Attorney General or any officer of the Department of Justice . . . establishes or implements a formal or informal policy to refrain . . . from enforcing, applying, or administering any provision of any Federal statute . . . whose enforcement, application, or administration is within the responsibility of the Attorney General or such officer on the grounds that such provision is unconstitutional." Subsection (e) of that statute extends this reporting obligation to the head of each executive agency or military department that implements such a policy of constitutional noncompliance. Such a report must be made within 30 days after the policy is implemented, and must "include a complete and detailed statement of the relevant issues and background (including a complete and detailed statement of the reasons for the policy or determination)." (The statute includes provisions for redaction of classified information in copies of the report distributed outside a congressional committee or agency.)

The Bush Administration, however, considers this reporting requirement itself to be constitutionally dubious, and thus appears to be refusing to enforce it, too. In his 2002 signing statement to the law containing the latest iteration of the provision, the President wrote:

Section 202 of the Act adds a new section 530D to title 28, United States Code, that purports to impose on the executive branch substantial obligations for reporting to the Congress activities of the Department of Justice involving challenges to or nonenforcement of law that conflicts with the Constitution. The executive branch shall construe section 530D of title 28, and related provisions in section 202 of the Act, in a manner consistent with the constitutional authorities of the President to supervise the unitary executive branch and to withhold information the disclosure of which could impair foreign relations, the national security, the deliberative processes of the Executive, or the performance of the Executive's constitutional duties. To implement section 202(b)(3) of the Act, the Attorney General, on my behalf, shall advise the heads of executive agencies of the enactment of section 202 and of this direction concerning construction of that section and section 530D of title 28.

This reluctance to let Congress and the public know how the Administration is, or is not, implementing federal law, is regrettable. As we have written elsewhere, transparency of the Executive’s legal judgments “helps to ensure executive branch adherence to the rule of law and guard against excessive claims of executive authority[, and] also promotes confidence in the lawfulness of governmental action.”

The final, and most important, problem with the practice in this Administration, as we emphasized above, is not the signing statements themselves, nor the simple fact that the President might be engaged in nonenforcement, but instead the substance of many of the Administration's constitutional objections: e.g., the extremely broad theories of the Commander-in-Chief Clause and the "unitary executive" that underlie many of the signing statements and other distorted statutory constructions. If those constitutional objections were well-taken, and were publicly disclosed and debated, the signing statements themselves would present far fewer problems. But many of us believe that the Administration is wrong on the merits; and it is that substantive concern, along with the concerns about the lack of transparency and about the use of nonenforcement as a tactic of first resort, that should be at the heart of this debate.

Unfortunately, the ABA Task Force Report has glossed over this important question of the substantive meaning of Executive prerogatives. In what seems like a misguided effort to fulfill its goal of nonpartisanship, the Report states “[o]ur recommendations are not intended to be, and should not be viewed as, an attack on the current President.” But criticism of a particular President’s abuse of power, far from being partisan, is an obligation of the legal profession.