A measure that has passed the House of Representatives and the Senate in identical form arrives, at presentment, at a fork that a single constitutional sentence built in advance. Article I, Section 7 of the Constitution assigns every presented bill one of a small set of endings, and the assignment depends on acts and conditions that the sentence specifies with unusual completeness. Signature makes the measure law. Return with objections sends the measure back to the chamber where the measure originated, where the objections enter the journal at large and reconsideration begins under a threshold of two thirds. Silence produces opposite results on opposite conditions, because inaction across ten days, Sundays excepted, enacts a bill that Congress could have received back and defeats a bill that adjournment placed beyond return. The mechanism decides outcomes before any argument about a particular measure begins, since the clock, the journal, and the threshold are fixed features of the procedure rather than terms open to negotiation in a given contest.

The fork matters because the presidential negative under Section 7 is a qualified negative rather than an absolute one. A returned bill carries a path back to enactment inside the same provision that authorizes the return, and that path runs through recorded votes in both chambers, with the names of the persons voting for and against the measure entered on the journal of each chamber respectively. The Constitution therefore builds accountability into the act of enactment over objection. A member who votes to pass a bill over a veto votes in public, by name, on a question framed by stated presidential objections that the originating chamber has placed on the permanent record. The design raises the price of override in visibility as well as in numbers, and the historical record reflects that combined price, because overrides have remained rare across more than two centuries of regular vetoes even in periods of open conflict between a president and a congressional majority.

The central claim of the account that follows concerns where the mechanism does most of its work. The threat does the work. A veto cast at presentment settles one measure, but a veto threatened during drafting, committee consideration, and negotiation between the chambers can reshape many measures without a single return taking place. The reason lies in the arithmetic of the override threshold. Drafters who count votes before passage count against two thirds of the members present and voting in each chamber, not against a simple majority, whenever a presidential objection has been stated in terms specific enough to credit. A provision that commands a comfortable majority and falls short of that higher margin sits inside the zone where a threatened return can force revision, deletion, or abandonment long before enrollment. Bargaining under the veto therefore takes place in the shadow of a vote that most participants prefer to avoid testing, and the bills that reach presentment in signable form often record, in their amended text, objections that were answered during construction rather than after return.

That shadow explains a paradox in the public record of the institution. The veto appears in constitutional argument as a dramatic confrontation between branches, complete with messages, journals, and roll calls, yet the ordinary operation of the power is quiet and anticipatory. Statements of administration policy fix a bargaining position while a measure is open to amendment. Packaging decisions change the leverage on both sides, because a provision opposed by a president travels differently alone than inside a large measure carrying provisions that a president seeks. Silence at presentment, the branch of the mechanism least visible of all, can enact a law without a signature or defeat a measure without a message, depending entirely on whether adjournment prevented return during the counted days.

The sections that follow examine the mechanism as a working system rather than as a civics diagram, from the constitutional text and the pocket veto litigation to the line item veto experiment and the record of overrides and signing statements. The through line holds across every stage. Enactment over objection is possible by design, difficult by design, and most influential, by design, at the moments when no veto has been cast at all.

The Constitutional Text and the Presentment Mechanism

The presidential negative on legislation has a precise textual home, and the location carries meaning before any dispute over a particular bill begins. Article I, Section 7 of the United States Constitution contains the presentment provisions, the clauses that govern the passage of a bill from congressional approval to legal force. Article I is the article that vests the legislative powers in Congress, organizes the House of Representatives and the Senate, regulates elections and meetings, sets rules for each chamber, restricts certain state and federal practices, and defines the method by which measures move through the legislative process. Placement inside that article places the presidential action at the end of the lawmaking sequence rather than in a separate catalogue of executive authority. Article II addresses the executive power, the term and election of the president, the commander in chief role, appointments, treaties, and the duty to take care that the laws be faithfully executed. Section 7 of Article I addresses a different moment, the moment at which a text approved by two chambers acquires or fails to acquire the status of law.

That placement reflects a structural choice about the character of the act. A signature under Section 7 completes a legislative process that began with introduction, committee consideration, debate, amendment, and passage in each chamber. A return under Section 7 sends the measure back into that same legislative process for reconsideration. In either direction, the presidential decision operates within the procedure for making law, and the constitutional text treats the decision as a stage of enactment rather than as an external act of administration directed at a law already in force. The distinction matters for interpretation because the powers and limits in Section 7 attach to the bill in its presented form, at a defined point, under a defined clock, and with defined consequences for silence, approval, and disapproval. The clause supplies the whole mechanism in compact form, and every later practice concerning signatures, returns, objections, journals, and reconsideration traces authority to that compact grant rather than to a general executive discretion.

The sequence that reaches presentment begins long before a bill reaches the president, and the account of how a bill becomes law in practice traces the stages of drafting, referral, floor action, and resolution of differences between the chambers that produce a single agreed text. Presentment has meaning only after that work is complete. A bill that passes one chamber in one form and the other chamber in another form has not satisfied the condition stated in Section 7, because the condition is passage by both the House and the Senate. The two chambers must approve the same measure, and the measure presented is the measure that both chambers passed. Presentment therefore marks the boundary between bicameral agreement and presidential consideration. Before presentment, the text belongs entirely to the legislative branch and congressional rules govern amendment and debate. After presentment, the text is fixed for purposes of the presidential decision, and the Constitution rather than chamber rules or presidential preference defines the available responses and the time allowed for a response.

Enrollment belongs to that boundary as a matter of form and authentication. After final passage, the bill is prepared in its enrolled form, the authoritative copy that embodies the text as passed and that is transmitted for presentment. Enrollment serves an evidentiary function in a system that depends on exact language. Statutes impose duties, authorize spending, define offenses, and create rights through specific words, and a discrepancy between the text passed and the text presented would place the validity of the resulting law in doubt. The enrolled bill resolves that risk by presenting one verified text as the record of bicameral action. Delivery of that enrolled bill to the president then performs a second function. Delivery is the act of presentment itself, and presentment starts the constitutional period for presidential action. The ten day period does not run from the last floor vote in either chamber, and the period does not run from enrollment as a clerical act. The period runs from presentment, from the point at which the passed bill is placed before the president for approval or disapproval under Section 7.

The definition of presentment in Section 7 reaches beyond instruments labeled as bills. A separate sentence in the same section provides that every order, resolution, or vote to which the concurrence of the Senate and the House of Representatives may be necessary, except on a question of adjournment, shall be presented to the president and shall require presidential approval, or, after disapproval, repassage by two thirds of the Senate and the House, before taking effect. The companion rule prevents evasion of presentment through a change of label. A joint action of the two chambers that requires concurrence and that is intended to have operative effect cannot escape presidential consideration or the override procedure merely because Congress styled the action as a resolution rather than as a bill. The exception for adjournment fits the same logic in reverse. Adjournment concerns the internal sitting arrangements of the chambers, and the Constitution excepts questions of adjournment from the presentment requirement that applies to concurrent orders, resolutions, and votes with substantive effect.

The ten day rule gives presentment its temporal force. Section 7 provides that if the president does not return a bill within ten days after presentment, Sundays excepted, the bill becomes law in the same manner as if the president had signed the bill, unless a stated exception applies. The exception for Sundays is written into the text and alters the count in every case in which the period crosses a Sunday. The count is therefore a constitutional count rather than a count supplied by statute, regulation, or custom. Ten days under Section 7 means ten days counted by the rule the clause itself supplies. The specificity of the count reflects the importance of certainty at the final stage of enactment. Interested persons, executive agencies, courts, and the chambers themselves require a determinate answer to the question of legal status, and an open ended period for presidential consideration would leave enacted texts in suspension and would invite dispute over the date on which obligations began.

Within the ten day period, the Constitution assigns the president three courses of action, with a fourth outcome produced by congressional adjournment in a defined circumstance. The first course is approval by signature. Section 7 states the sequence in direct terms. Every bill passed by the House and the Senate shall, before the bill becomes law, be presented to the president. If the president approves, the president signs the bill, and the bill becomes law. Signature is an affirmative act performed on the presented measure, and legal force follows from that act together with the prior passage by both chambers. Signature does not amend the text, add conditions to the text, or select portions of the text for approval. The presented bill is the unit of action. The president approves the bill as passed or withholds approval from the bill as passed, and the signature, when given, attests approval of the whole measure that presentment placed before the presidency.

The second course is return with objections, the course commonly called a regular veto in later practice and in historical counting. If the president disapproves a bill, Section 7 directs return of the bill with objections to the chamber in which the bill originated. The originating chamber is the House if the bill began in the House and the Senate if the bill began in the Senate. Return is therefore directed rather than general. The objections travel with the bill to a specified destination, and the Constitution then assigns duties to that destination. The originating chamber enters the objections at large on its journal and proceeds to reconsider the bill. Entry at large means entry of the objections in full on the official record of the chamber, rather than entry by summary or by reference alone. The journal entry preserves the stated grounds of disapproval in the legislative record and makes those grounds available to members during reconsideration, to the other chamber during its own reconsideration, and to readers of the record after the controversy has passed.

Return with objections differs from silence, from informal criticism, and from a public statement of displeasure unaccompanied by return of the bill. The constitutional act consists of return of the presented bill together with objections within the period allowed. Objections are a required part of the act rather than an optional explanation attached to the act. A president who disapproves must state objections, and the text directs those objections to the chamber that first passed the bill. The requirement gives reconsideration a defined subject. Members of the originating chamber vote again on a known text for stated reasons, and debate on reconsideration can address the text, the objections, or the relationship between the objections and the text. The requirement also disciplines the executive side of the exchange. Disapproval must be explained in a form that enters the journal, and a bare refusal without stated grounds does not satisfy the procedure that Section 7 prescribes for a return.

The third course is inaction within the period, and Section 7 assigns inaction a legal consequence equal in force to signature in the ordinary case. If the president neither signs nor returns a bill within ten days, Sundays excepted, the bill becomes law as if signed. The clause removes any pocket of uncertainty in which a bill passed by both chambers and presented in regular course would fail through presidential silence alone. Silence after presentment, with Congress in circumstances that permit return, produces enactment rather than defeat. The rule places the burden of disapproval on an affirmative act. A president who opposes a presented bill and who seeks to prevent enactment through the regular path must return the bill with objections in time for reconsideration. Failure to perform that act within the period allows the bill to take effect without a signature, and the absence of a signature in that circumstance carries no inference of invalidity, because the text itself supplies the rule that converts inaction into enactment.

The fourth outcome, the pocket branch, arises only when the stated exception displaces the ordinary consequence of inaction. Section 7 provides that a bill unsigned and unreturned after ten days does not become law if Congress by adjournment has prevented return of the bill within that period. The exception turns on prevention of return, and prevention of return turns on adjournment by Congress. In that circumstance, the bill fails without a return, without objections entered on a journal, and without an opportunity for reconsideration and an override vote in that session on that presented bill. The branch is therefore structurally distinct from a regular return. A regular return opens a further legislative stage, reconsideration under a heightened voting rule. The pocket branch closes the process at presentment because the physical and procedural act of return has been prevented by adjournment. Later sections of analysis address the case law and practice that developed around the meaning of adjournment and prevention, and the text alone supplies the starting point for that analysis in the conditional language of Section 7.

The decision structure can be stated as a standing sequence without any resort to a chart or catalogue. Presentment occurs after passage by both chambers. The ten day period, Sundays excepted, begins at presentment. Signature within the period produces law. Return with objections within the period sends the bill to the originating chamber, places the objections on the journal of that chamber, and triggers reconsideration. Inaction through the end of the period produces law as if signed, provided that adjournment has not prevented return. Adjournment that prevents return within the period produces failure of the bill without signature, without return, and without reconsideration. Each branch depends on an act or condition that the text names, and no branch depends on presidential silence being interpreted as approval or disapproval in a psychological sense. The Constitution assigns legal consequences to observable events, signature, return, lapse of the period, and adjournment preventing return, and the assignment leaves no branch to inference from motive.

A central feature of the mechanism is the qualified character of the presidential negative. Section 7 does not grant an absolute power to nullify a bill passed by both chambers. A return begins reconsideration, and reconsideration can end in enactment over the stated objections if the required majorities agree in each chamber. The negative is qualified by the override rule stated in the same clause, and the grant and the qualification appear together rather than in separate provisions that might be read in isolation. Analysis that quotes the return language without the reconsideration language therefore misstates the power. The text authorizes the president to require a second consideration under a higher threshold, to place objections on the record as the basis for that consideration, and to defeat the bill if the higher threshold is not met in either chamber. The text withholds authority to defeat the bill conclusively by return alone when both chambers meet the higher threshold in sequence.

Language reinforces the point about qualification. The word veto does not appear in the Constitution. Section 7 uses the verbs of the process, present, approve, sign, return, and reconsider, and the noun objections for the statement that accompanies a return. Veto entered usage as a convenient label for the return power and for the message that exercises the power, and the label is serviceable as shorthand in historical writing and in ordinary discussion. Shorthand carries a risk, however, because the Latin derived label can suggest a flat prohibition of the kind associated with an absolute negative in other institutional settings. The constitutional mechanism is narrower and more dialogic than the label alone suggests. The president returns a bill, states objections, and submits the measure to a prescribed second judgment by the elected chambers. Precision about vocabulary therefore serves precision about power, and a reader who follows the verbs in Section 7 receives a more accurate account of the institution than a reader who relies on the single imported label.

The objections requirement also connects the return power to legislative deliberation rather than to executive command. Objections entered at large on the journal become part of the materials of reconsideration, and the originating chamber proceeds to reconsider the bill in light of the stated grounds. The grounds may be constitutional, prudential, fiscal, administrative, or grounded in policy disagreement, because Section 7 does not limit the permissible content of objections to a single category. The text requires objections and prescribes the destination and the record, and the text leaves the substantive argument open. That openness has a structural consequence. A return can challenge the constitutionality of a provision, the wisdom of a program, the cost of an authorization, or the drafting of a definition, and each form of challenge enters the same procedural channel and faces the same reconsideration rule. The channel is uniform even though the arguments vary, and uniformity of channel is part of the design that Section 7 establishes for conflicts between the branches at the point of enactment.

Journal entry and recorded voting extend the same concern for a public and verifiable record. Section 7 directs the originating chamber to enter the objections at large on its journal, and the clause also provides that, in override cases, the votes of both houses are determined by yeas and nays and the names of persons voting for and against the bill are entered on the journal of each house respectively. Recorded voting by name is therefore a constitutional requirement for reconsideration after return rather than a courtesy adopted by chamber rule. The requirement raises the visibility of the decisive stage. A member who votes to enact a returned bill over objections casts a recorded vote on a measure that the president has publicly disapproved for stated reasons, and a member who votes against enactment likewise casts a recorded vote in support of the stated objections or against the bill on other grounds. The journal preserves both the reasons and the names, and preservation serves accountability in elections, accuracy in historical accounts, and certainty in proof of enactment.

The reconsideration rule that follows a return is stated with a precision that later practice has interpreted in a specific sense. Section 7 provides that if, after reconsideration, two thirds of the originating house agree to pass the bill, the bill is sent, together with the objections, to the other house, by which the bill is likewise reconsidered, and if approved by two thirds of that house, the bill becomes law. Congressional procedure, tradition, and judicial ruling interpret the phrase two thirds of that house as two thirds of members present and voting, a quorum being present. The distinction between that interpretation and a requirement of two thirds of all members is material in close cases, because vacancies, absences, and abstentions affect the denominator under a present and voting standard in a manner different from a full membership standard. The interpretation belongs to the meaning of the text as applied, and later discussion of individual override votes depends on the denominator being understood correctly from the outset.

Understanding the mechanism at the textual level therefore requires attention to five elements held together in one clause. The first element is location, Article I, Section 7, inside the lawmaking article. The second element is presentment, delivery of the passed and enrolled bill to the president, together with the companion rule for orders, resolutions, and votes requiring concurrence, adjournment excepted. The third element is time, ten days, Sundays excepted, running from presentment. The fourth element is choice among defined acts, signature, return with objections to the originating chamber for journal entry and reconsideration, or inaction with the consequence the text assigns. The fifth element is qualification, a negative subject to repassage by two thirds of members present and voting in each chamber, with votes by yeas and nays and names entered on the journals. Together, the five elements form the constitutional foundation on which the history of individual vetoes, overrides, pocket disputes, and related practices rests, and the foundation is stated in the text with an economy that later practice has elaborated without displacing.

What does enrollment do before a bill is presented to the president?

Enrollment prepares the authoritative copy of a bill after final passage, embodying the exact text that both chambers passed. That verified copy is then transmitted for presentment. Delivery of the enrolled bill is presentment itself, and delivery starts the constitutional period allowed for presidential approval or return with objections.

The Mechanics in Operation Step by Step

The sequence begins at a defined threshold. A bill passed by the House of Representatives and the Senate does not become law on passage alone. Before a bill becomes law, the measure is presented to the president. Presentment is the act that sets the rest of the machinery in motion, because the constitutional choices that follow are choices about a measure that has completed passage in both chambers and has then been placed before the president for approval or disapproval. The same presentment rule reaches orders, resolutions, and votes that require the concurrence of both chambers, with adjournment excepted from that class, and the operation described in this section concerns the treatment of a bill after that threshold has been crossed. Nothing in the ten day period runs before presentment, and nothing in the reconsideration procedure begins before a return has taken place. The order of the steps is therefore part of the substance. Passage comes first, presentment follows, the presidential decision follows presentment, and congressional reconsideration follows only a return with objections.

The ten day clock starts at presentment. From that point the president holds the measure under a period measured in days, and the Constitution assigns a consequence to each way in which the period can end. Signature within the period makes the measure law. Return within the period, with objections, sends the measure back to Congress for reconsideration. Inaction across the whole period, while Congress sits in a position that permits return, also makes the measure law, as if the president had signed it. Inaction across the whole period, where adjournment by Congress prevents return within the period, produces the opposite result, and the measure is not law. The clock is therefore not a waiting period with a single meaning. The clock is a sorting device. The same lapse of ten days carries opposite legal consequences depending on whether return was possible during those days, and the same act of signing or returning carries its consequence only because it occurs within the frame that presentment opened.

Sundays are excepted from the count. The period is ten days with Sundays left out, so the count moves through the ordinary days of the week and steps over each Sunday that falls inside the period. A measure presented shortly before a Sunday does not lose that Sunday against the period, and a period that spans two Sundays leaves both Sundays out of the reckoning. The exception lengthens the calendar span without lengthening the counted span, and it does so by a fixed rule rather than by discretion in the particular case. No officer extends the period because business was heavy, and no officer shortens the period because a decision was reached in draft before the period expired. The rule treats every presented measure alike. The days are counted, Sundays are excepted, and at the end of the counted days the branch that governs the measure is determined by what the president did or failed to do and by whether return stood open during the period.

The signature branch is the most direct of the branches. The president approves the measure and signs it, and the measure becomes law. Signature requires no statement of reasons, no entry of objections, and no further action by either chamber, because the two chambers have already passed the bill in the form presented. The law that results from signature is the text that both chambers passed and that the president received at presentment. No stage after signature reopens the wording, and no stage after signature asks either chamber to pass the measure a second time. The contrast with the return branch defines the signature branch. Approval ends the sequence. Disapproval, expressed in the manner the Constitution prescribes, extends the sequence by one further congressional stage, and silence, under the conditions described below, ends the sequence in enactment without any signature at all.

The return branch begins with disapproval expressed as a return. If the president disapproves a bill, the bill is returned with objections to the chamber in which the bill originated. The returning measure travels unsigned, and the absence of a signature is matched by the presence of a written statement of the grounds of disapproval. A return without objections does not satisfy the described procedure, and a statement of objections without return of the bill does not satisfy the described procedure either. The bill and the objections move together, because the chamber that receives them is asked to reconsider the very measure that was returned, in light of the very objections that accompany the measure. The destination matters as much as the contents. Return goes to the originating chamber, the House of Representatives for a measure that began there and the Senate for a measure that began there, and reconsideration takes its order from that destination.

On arrival in the originating chamber, the objections are entered at large on the journal of that chamber. Entry at large puts the objections on the record in full rather than in summary or by reference, so the journal of the chamber carries the complete statement that the president sent with the returned bill. The entry precedes and accompanies reconsideration. Members acting on the returned measure act on a measure whose stated objections stand spread on the record of their own chamber, and the later stages of the procedure carry the same pairing forward. When the bill passes from the first chamber to the second after reconsideration, the bill goes with the objections, so the second chamber also reconsiders a measure accompanied by the full statement of the case against the measure. At no point in the sequence do the objections drop away and leave a bare bill to be voted on as though the return had never occurred. The returned bill and the objections form a single package from the moment of return to the end of reconsideration in the second chamber.

Reconsideration begins in the originating chamber. The chamber that first passed the bill in its original passage is also the chamber that first takes up the bill after return, and the order is fixed rather than chosen case by case. The originating chamber proceeds to reconsider the bill after the objections have been entered at large on its journal. Reconsideration in this setting has a specific object. The chamber does not draft a substitute measure, does not amend the returned text through the ordinary amendment process as a step in the override sequence, and does not take up a fresh bill on the same subject in place of the returned bill. The chamber reconsiders the returned bill itself, the identical measure that passed both chambers, went to the president at presentment, and came back unsigned with objections. A chamber that wishes to change the text must look outside the override sequence for that purpose, because the override sequence answers a narrower question about the returned measure in its returned form.

The question put in reconsideration is framed by the objections. The chamber votes on passing the bill notwithstanding the objections of the president. That framing shapes the debate and the vote alike, because a vote in favor is not recorded as a simple repetition of the original vote for passage. A vote in favor declares that the bill ought to pass even after the stated objections have been entered, read, and weighed. A vote against declares that the bill ought not to pass in the face of those objections, whether the voter accepts the objections on the merits, accepts part of the reasoning, or holds that the measure lacks the heightened support the procedure demands. The heightened demand is built into the same step. Agreement of two thirds of that chamber passes the bill on reconsideration, and a simple majority, sufficient for original passage, does not suffice at this stage. The measure that attracted a majority on first passage must attract two thirds on reconsideration before the measure moves any further along the override path.

The two thirds requirement carries a settled measure of its own. Two thirds of that chamber means two thirds of the members present and voting, a quorum being present, as congressional procedure, tradition, and judicial ruling have interpreted the constitutional phrase. The requirement is never a requirement of two thirds of all members of the chamber. The distinction decides real cases, because attendance on a given day determines the number against which the two thirds share is calculated. A chamber acting with a quorum present acts through the members present and voting, and agreement by two thirds of those members satisfies the constitutional demand at that stage. The Supreme Court sustained that reading in Missouri Pacific Railway Co. v. Kansas, 248 U.S. 276 (1919), a unanimous decision in an opinion by Chief Justice White, where two thirds of a quorum present was held sufficient for an override in the challenge to the enactment of the Webb-Kenyon Act over a veto. The case fixed the denominator as a matter of constitutional interpretation, and the procedure described in this section follows that fixed reading at both reconsideration stages.

The vote itself is also prescribed in form. In all cases of reconsideration after a return, the votes of both chambers are determined by yeas and nays, and the names of the persons voting for the bill and the names of the persons voting against the bill are entered on the journal of each chamber respectively. The recorded vote requirement stands written into the constitutional text itself, rather than resting on a rule that either chamber adopted and either chamber could waive. Each member voting on reconsideration therefore votes on the record by name, and the journal of the chamber preserves the division of names after the vote has closed. The requirement applies in the originating chamber and applies again in the second chamber, so a completed override leaves two recorded divisions, one in each journal, with the names arranged on each side of the question. A voice vote, a division without names, or any unrecorded form of decision does not meet the prescribed form for this stage of lawmaking.

Passage in the originating chamber sends the sequence onward. If two thirds of that chamber agree to pass the bill notwithstanding the objections, the bill is sent, together with the objections, to the other chamber. The transmission repeats the pattern set at the return. The bill travels with the statement that prompted reconsideration, and the second chamber receives both the measure and the grounds on which the president withheld approval. The second chamber then reconsiders the bill in the same form as the first. The question in the second chamber is again passage notwithstanding the objections, the measure under consideration is again the identical returned bill, and the threshold is again two thirds of the members present and voting, a quorum being present, with the vote taken by yeas and nays and the names entered on the journal of that chamber. Approval by two thirds of the second chamber completes the enactment, and the bill becomes law through the agreement of two thirds in each chamber in succession, without a signature and against the stated objections that started the reconsideration sequence.

The order of the two reconsiderations carries consequences for the fate of the measure. The second chamber takes up the bill only after the originating chamber has passed the bill on reconsideration, because the text sends the bill to the second chamber after agreement in the first. A failure to reach two thirds in the originating chamber ends the override sequence at that point, and the returned bill does not pass. A failure to reach two thirds in the second chamber, after passage in the first, likewise ends the sequence without enactment, because enactment through reconsideration rests on approval by two thirds of both chambers. Neither chamber acts for the other, and neither recorded division substitutes for the other. The procedure demands two separate heightened majorities, assembled in sequence, in two distinct chambers, on two distinct journals. That doubleness is the price the Constitution sets for enacting a measure over a presidential return, and the price is paid in the prescribed order or the measure fails at the stage where the required agreement was not reached.

The branch of law without signature occupies the ground between signature and return. If the president neither signs a presented bill nor returns the bill with objections within ten days, Sundays excepted, the bill becomes law in like manner as if the president had signed it, subject to the single exception treated in the paragraphs that follow. Inaction, under the stated conditions, enacts. The branch removes any need for an affirmative act at the end of the period, and the legal effect at the close of the tenth counted day matches the legal effect of a signature placed on the measure during the period. The text passed by both chambers becomes law whole, in the form presented, with no objections entered on any journal and no reconsideration held in either chamber. Members who supported the bill gain enactment without a second vote, members who opposed the bill face a law enacted without a presidential signature, and the record of the sequence shows presentment followed by the lapse of the full period with the measure neither signed nor returned.

That branch depends on a condition that the wording of the rule makes central. Inaction enacts only where return stood open during the period. A president who holds a bill unsigned for ten days while Congress sits in a position to receive a return has, in legal effect, allowed the period to run with every branch available. Signature was available and was not used. Return was available and was not used. The enactment that follows rests on that availability, because the president had a working path for sending the bill and the objections back to the originating chamber at any point during the counted days and let the days expire without taking that path. Congress, for its part, stood ready to receive a return, to enter objections at large on a journal, and to begin reconsideration, and that readiness was never called on. Law without signature is therefore best understood as enactment after a declined opportunity to return, rather than enactment after a blocked opportunity to return. The distinction between a declined opportunity and a blocked opportunity governs the final branch of the period.

The adjournment branch states the exception with precision. Unless Congress by adjournment prevents the return of the bill within the ten day period, inaction enacts, and where Congress by adjournment does prevent return within the period, the bill is not law. Adjournment matters only where adjournment prevents return. An adjournment that leaves return possible does not trigger the exception, an adjournment that begins after the period has closed does not trigger the exception, and a sitting pattern that keeps a path of return open throughout the counted days leaves the ordinary rule of inaction untouched. The test fastens on prevention during the period, on the practical and legal capacity of the originating chamber to receive back its bill with objections before the tenth counted day expires. Where that capacity was absent because Congress had adjourned in a manner that prevented return, silence by the president cannot be treated as acquiescence in enactment, because the president never held the full choice that the inaction branch presupposes. A measure that fails under this branch is not law, and no override is possible in that posture, because nothing was returned, no objections were entered on a journal, and no reconsideration sequence was ever set in motion.

Read together, the four outcomes of the ten day period form a complete and closed set. Signature enacts the measure at once. Return with objections suspends the fate of the measure and transfers the decision to Congress under the heightened two thirds rule in each chamber, acting in sequence and voting by yeas and nays on the record. Inaction while return stands open enacts the measure at the close of the period as if signed. Inaction where adjournment prevents return defeats the measure at the close of the period, with no path into reconsideration. Each outcome turns on acts and conditions inside the same counted span, Sundays excepted, that began at presentment. The president holds the choice between signature and return throughout the span. Congress holds the condition that decides the meaning of silence, because the sittings and adjournments of Congress determine whether return stood open while the days ran. Neither side controls the whole sequence alone, and the distribution of control across the period is the operating character of the qualified negative that the Constitution establishes. The veto is a qualified negative rather than an absolute one, since a returned measure can become law on the agreement of two thirds in both chambers, and the word veto itself appears nowhere in the constitutional text that creates the power.

The first exercise of the power followed the return branch from beginning to end in its opening stages. On April 5, 1792, President George Washington returned the Apportionment Act unsigned, in the first regular veto of a measure under the Constitution. The objections that accompanied the returned bill rested on constitutional grounds. Washington objected that the measure provided no uniform ratio for representation across the states, and Washington objected that eight states had been allotted more than one representative for every thirty thousand inhabitants. The objections were entered for the originating chamber to weigh on reconsideration, under the same sequence of return, journal entry, and reconsideration that governs every later return. That first return set the template in practice for the mechanics described in this section, a presented bill withheld from signature, sent back with stated constitutional objections to the chamber where the bill began, and placed before Congress again on the question of passage notwithstanding the objections.

Which measures besides bills must also be presented to the president?

The presentment rule also reaches every order, resolution, or vote that requires the concurrence of both the House and the Senate, except a question of adjournment. Such measures follow bills to the president because legal effect for them likewise depends on action completed jointly by the two chambers.

The Presentment Decision Tree

The four presidential options under Section 7 differ less in the acts available than in the interaction between each act and the calendar. Signature and return are affirmative choices made inside the same counted period, and inaction takes its legal meaning entirely from the condition of Congress during that period, so a comparison that separates the choice from the timing misstates the mechanism. The table sets each option against the timing that governs the option, the consequence for the measure, and the response open to Congress, using only the outcomes that the constitutional text and the practice under the text establish.

Presidential option	Timing	What happens to the measure	Congressional response
Signing the bill	Within ten days, Sundays excepted, after presentment	The measure becomes law	No response is required, and the enacted text enters the statute book
Returning the bill with objections	Within ten days, Sundays excepted, after presentment	The bill returns to the chamber where the bill originated, the objections are entered at large on the journal of that chamber, and the chamber proceeds to reconsider the bill	Two thirds of the members present and voting in the originating chamber, a quorum being present, passes the bill to the other chamber, and two thirds of the members present and voting there enacts the bill, with votes by yeas and nays and names entered on the journal of each chamber
Taking no action while Congress can receive a return	The ten day period, Sundays excepted, expires with return possible	The bill becomes law as if the president had signed the bill	No response is required, and enactment rests on the expiration of the period rather than on signature
Taking no action where adjournment prevents return	The ten day period, Sundays excepted, expires with adjournment preventing return	The bill does not become law, in the outcome known as a pocket veto	No override is possible, because no bill and no objections were returned for reconsideration, and enactment requires passage of the measure again

The Override Threshold, the Recorded Vote, and the Odds

The override threshold is the point at which the veto stops being a presidential decision and becomes a test of congressional resolve measured against a demanding constitutional standard. Article I, Section 7 provides that a returned bill passes on reconsideration by agreement of two thirds of that House, that the bill then goes with the president’s objections to the other chamber, and that approval by two thirds of that chamber enacts the bill into law. The operative constitutional phrase is two thirds of that House. Congressional procedure, tradition, and judicial ruling read that phrase with a precision that casual summaries often lose. The requirement is two thirds of the members present and voting in each chamber, a quorum being present. The denominator is the body of members who take part in the vote under conditions that satisfy the quorum requirement. The denominator is not the full elected membership of the chamber, and it does not expand to include members who are absent, members who abstain from voting, or seats that stand vacant at the time of the vote. That distinction governs every serious account of the override power, because a threshold described with the wrong denominator misstates the size of the coalition that the Constitution demands and misleads any reader who attempts to judge how difficult a particular override was to assemble or how close a failed attempt came to success.

The common misstatement that an override requires two thirds of all members must be named as an error and corrected wherever the mechanism is explained. Two thirds of all members would be a fixed number tied to the total membership of each chamber, indifferent to attendance and indifferent to participation. The constitutional rule, as procedure, tradition, and judicial ruling have settled the reading, operates differently. A member who is present and voting counts in the denominator and, by voting against reconsideration, raises the number of affirmative votes required. A member who is absent does not count in the denominator at all. A member who is present but does not vote likewise falls outside the count of members present and voting, although the presence of that member may contribute to the quorum that makes the proceeding valid. The distinction between a quorum and a vote is therefore central to the threshold. A quorum being present establishes that the chamber is competent to act. The votes cast by members present and voting then supply the measure against which the two thirds fraction is calculated. Collapsing the two concepts into a single reference to total membership produces an account that sounds stricter than the Constitution and that fails to describe the vote that either chamber actually conducts.

The Supreme Court settled the reading of the phrase in Missouri Pacific Railway Co. v. Kansas, 248 U.S. 276 (1919). The case challenged the enactment of the Webb-Kenyon Act, a statute that had become law over a presidential veto, and the challenge turned on whether the override vote in question had satisfied the constitutional requirement. The contention pressed against the statute treated the phrase two thirds of that House as though it demanded support measured against a denominator broader than the members who had participated. The Court rejected that reading. In a unanimous decision, with the opinion associated with Chief Justice White, the Court held that two thirds of a quorum present suffices for an override. The holding confirmed the interpretation that congressional procedure and tradition had applied. Two thirds of the members present and voting, a quorum being present, satisfies the Constitution. Two thirds of all members is not the test, and an enactment completed over a veto by the lesser but constitutionally sufficient margin stands as law. The decision gave the settled practice the authority of judicial ruling, so that the threshold rests on three mutually reinforcing foundations rather than on habit alone.

The recorded vote requirement stands beside the threshold in the same constitutional passage, and the two provisions operate as a pair. In all override cases, the votes of both Houses are determined by yeas and nays, and the names of the persons voting for and against the bill are entered on the journal of each House respectively. The requirement is mandatory in form. An override cannot be completed by an unrecorded voice vote, by a division in which individual positions go unrecorded, or by any procedure that leaves the identity of supporters and opponents outside the official record. Each chamber must take the vote in a manner that assigns a position to a name, and each chamber must enter those names on its own journal. The first chamber to reconsider creates a record that accompanies the bill and the objections to the second chamber, and the second chamber, in its turn, creates a record of equal specificity. By the time a returned bill becomes law over a veto, or fails on reconsideration, the constitutional process has generated two named tallies that together show exactly which members sustained the president and which members voted to enact the bill despite the objections returned with it.

That requirement holds a distinctive place in the constitutional text because it is one of very few recorded vote requirements written into the Constitution itself. Much of congressional procedure rests on rules that each chamber adopts, on statutes that structure legislative business, and on practices that developed through precedent. A chamber can ordinarily amend its own rules, and a practice can evolve as majorities and circumstances change. The yeas and nays requirement for override votes admits no such revision. The command sits in Article I, Section 7, in the same sentences that create the return and reconsideration process, and neither chamber can dispense with the recorded vote by rule, by unanimous consent, or by convenience. The text treats the manner of voting as part of the substance of the override power rather than as a housekeeping detail left to later arrangement. Because the requirement appears in the Constitution, a failure to record the votes by name would not be a mere irregularity in internal procedure. Such a failure would depart from the process by which the Constitution authorizes a bill to become law without presidential approval, and the override route is the only route the text provides for enactment over a returned veto.

The accountability function of the recorded vote explains why the text insists on names. A veto override is a direct institutional confrontation between the legislative branch and the executive branch over a measure that both branches have considered in full. The president has stated objections on the record. Each chamber has already passed the bill once, and reconsideration asks each member to decide whether the original judgment should prevail against the objections of the president who returned the measure. A recorded vote attaches a public position to every member who participates, at the moment of the most visible disagreement the lawmaking process produces. Constituents can read the journal entry and learn how a representative or a senator voted when the choice was framed in its sharpest form, enactment against the stated objections of the president or deference to those objections after reconsideration. Political consequences, whether supportive or adverse, can then follow from an identifiable act rather than from inference or collective attribution. The requirement also disciplines members during the vote itself. A position that will appear by name on the journal invites a deliberative seriousness that an anonymous tally cannot supply, and a member who changes position between passage and reconsideration does so on a record that permits comparison between the two votes.

The odds of an override, stated at named source strength, give the threshold an empirical meaning that the constitutional text alone cannot supply. The source for the figures is the veto table of the Senate Historical Office, covering 1789 through the close of the 116th Congress, and the span and the source govern every figure that follows. Across 1789 through the close of the 116th Congress, the Senate Historical Office table records 2,584 vetoes in total. Of that total across 1789 through the close of the 116th Congress, 1,518 were regular vetoes, meaning vetoes accomplished by the return of a bill with objections to the chamber where the bill originated, and 1,066 were pocket vetoes. The same table, for 1789 through the close of the 116th Congress, records 112 overrides. Set against the correct denominator, those figures yield the standard measure of override frequency. One hundred twelve overrides out of 1,518 regular vetoes across 1789 through the close of the 116th Congress is about 7.4 percent, roughly one regular veto in fourteen. The statement of the rate, the identification of the source, and the identification of the span belong together. Detached from the Senate Historical Office table and detached from 1789 through the close of the 116th Congress, the percentage would claim an authority that no figure in this account possesses.

The denominator demands emphasis because an incorrect denominator produces a materially false rate. Pocket vetoes cannot be overridden. A pocket veto arises when adjournment prevents the return of a bill within the ten day period, Sundays excepted, and the bill does not become law. Nothing is returned, no objections are presented for reconsideration, and no chamber receives the occasion to vote on passage over a veto. The override procedure has no object on which to operate. Including 1,066 pocket vetoes from 1789 through the close of the 116th Congress in the denominator of an override rate would therefore count, as failures of reconsideration, a large class of cases in which reconsideration was constitutionally unavailable. The resulting figure would understate the frequency with which Congress enacted bills over returned vetoes and would misdescribe the institution by treating a structural impossibility as a political outcome. The ratio of about 7.4 percent applies to regular vetoes only, and the ratio must never be applied to all vetoes. Regular vetoes are the sole class of vetoes that present the question the override clause answers, and the population of 1,518 regular vetoes across 1789 through the close of the 116th Congress is the only population against which the 112 overrides in that span can honestly be measured.

Franklin D. Roosevelt issued the most vetoes of any president in the table covering 1789 through the close of the 116th Congress, with 635 vetoes in total, divided between 372 regular vetoes and 263 pocket vetoes, and 9 of those vetoes were overridden. The combination illustrates the protective effect of partisan alignment during much of a long presidency. Grover Cleveland, across two presidencies combined, issued 584 vetoes in total in the figures covering 1789 through the close of the 116th Congress, divided between 346 regular vetoes and 238 pocket vetoes, and 7 of those vetoes were overridden. The Cleveland figures show a president who relied on the veto, and especially on the pocket veto, at a volume approached by few others, while facing an override outcome in only a small fraction of cases. Gerald Ford presents the contrasting pattern in the distribution covering 1789 through the close of the 116th Congress. Ford issued 66 vetoes in total, divided between 48 regular vetoes and 18 pocket vetoes, and 12 of those vetoes were overridden. Twelve overrides against a total of 66 vetoes is a rate of successful reconsideration far beyond the aggregate pattern of about one regular veto in fourteen, and the Ford figures stand as the clearest illustration in the distribution of how the odds shift when a president faces Congresses in which the opposing alignment can plausibly assemble supermajorities. Ronald Reagan issued 78 vetoes in total in the figures covering 1789 through the close of the 116th Congress, divided evenly between 39 regular vetoes and 39 pocket vetoes, and 9 of those vetoes were overridden. The Reagan distribution sits between the Ford pattern and the patterns of Roosevelt and Cleveland. Vetoes were a regular instrument of policy, overrides were a recurring rather than exceptional outcome, and the division of control between the branches during the period gave each side victories that the aggregate rate, drawn from 1789 through the close of the 116th Congress, would not predict for either side in isolation.

The later presidencies in the distribution reinforce the lesson that alignment, rather than any fixed property of the veto itself, drives the outcome. George H. W. Bush issued 44 vetoes in total in the figures covering 1789 through the close of the 116th Congress, divided between 29 regular vetoes and 15 pocket vetoes, and 1 of those vetoes was overridden. Bill Clinton issued 37 vetoes in total across the same span of the table, divided between 36 regular vetoes and 1 pocket veto, and 2 of those vetoes were overridden. George W. Bush issued 12 regular vetoes and no pocket vetoes in the table covering 1789 through the close of the 116th Congress, and 4 of those vetoes were overridden, a proportion of successful reconsideration that ranks high against the aggregate rate even though the absolute number of vetoes was small. Barack Obama recorded one override in the table covering 1789 through the close of the 116th Congress, the override of the veto of the Justice Against Sponsors of Terrorism Act in 2016. Donald Trump recorded one override through the close of the 116th Congress, the override of the fiscal 2021 defense authorization veto, completed on January 1, 2021. Read together, the figures from these presidencies show small totals paired with consequential single outcomes, and the contrast between one override in a presidency and twelve overrides in another measures the distance between Congresses that could rarely reach the threshold and Congresses that could reach the threshold repeatedly. Subject matter could rearrange alignments that party ties ordinarily held together, because a bill touching interests with support across party lines could draw affirmative votes on reconsideration from members of the president’s own party. The distribution captures the cumulative effect of those interactions. The figures record outcomes, and explanation of the outcomes requires attention to the alignment and the measure in each case.

A counter reading of the evidence is required by any honest treatment of the subject, because the most available interpretation of an override is also the most careless one. An override is not by default evidence of presidential weakness. The conclusion does not follow from the event, and the structure of the threshold shows why the conclusion fails. Assembling two thirds of the members present and voting in the House of Representatives and two thirds of the members present and voting in the Senate usually indicates that a measure carried bipartisan support of an unusual breadth, support that the president opposed and could not fracture. A coalition of that scale is evidence about the measure and about the distribution of opinion in the legislative branch at least as much as evidence about the president. Presidents veto bills for reasons that include constitutional objection, policy disagreement, and the judgment that a measure with broad support is nevertheless unsound or unconstitutional. A president who vetoes a popular measure and loses the override vote has exercised the qualified negative exactly as the Constitution contemplates, and the loss records the resolution of a disagreement between the branches rather than a collapse of executive authority. The same event that one account describes as a defeat can be described, with equal fidelity to the record, as the constitutional process completing a cycle in which each branch performed the function assigned to the branch.

The executive check argument gives that counter reading its fullest form. On this account, the veto exists to give the president an independent defensive power against hasty, ill considered, or unconstitutional legislation, and the override exists to keep that power qualified rather than absolute. The president who returns a bill forces a second consideration under a higher threshold, compels the legislative branch to state its commitment at supermajority strength, and places objections on the journal record for both chambers and for the public to weigh. Even a successful override vindicates part of that design, because the law that emerges has survived a scrutiny that ordinary legislation never faces and has demonstrated support broad enough to satisfy the most demanding voting rule in the ordinary lawmaking process. From the executive perspective, the rarity of overrides across 1789 through the close of the 116th Congress, about one regular veto in fourteen on the figures of the Senate Historical Office table, shows a check that operated with restraint on both sides. Presidents returned bills that Congresses usually declined to enact over objection, and Congresses reserved successful reconsideration for measures whose support crossed the lines that ordinarily sustain a veto. The argument treats the override not as a repudiation of the presidency but as the outer boundary of a negotiation between branches that the veto initiates and that reconsideration, in a small minority of cases, concludes against the president.

The legislative supremacy argument, presented with equal care, reads the same record from the opposite institutional position. On this account, lawmaking authority rests with the elected legislature, the veto is an exception to legislative authority justified only as a qualified negative, and the override is the mechanism by which the legislative branch reasserts its primacy when a single executive officer blocks a measure that commands overwhelming support. The two thirds threshold, high as the threshold stands, marks the point at which the judgment of the representatives in both chambers outweighs the judgment of the president, and the recorded vote by name ensures that the reassertion occurs openly, member by member, on the journals of both Houses. From the legislative perspective, the 112 overrides recorded across 1789 through the close of the 116th Congress represent instances in which Congress corrected an executive obstruction of measures with exceptional support, and the distribution of those overrides across presidencies of differing alignments shows a power that no durable partisan arrangement permanently disabled. The argument also draws force from the threshold itself. A branch that must reach two thirds of the members present and voting in each chamber, on a recorded vote, after consideration of stated presidential objections, and that reaches the threshold nonetheless, has made the strongest showing of legislative will that the Constitution recognizes outside the amendment process. A balanced reading holds both accounts against the mechanics of the clause itself. The threshold of two thirds of the members present and voting, a quorum being present, as Missouri Pacific Railway Co. v. Kansas confirmed in 1919, makes the override difficult by design and possible by design. The votes by name requirement makes the attempt accountable by design. The outcome in any given case belongs to the alignment, the measure, and the coalition that the recorded vote reveals, and the long record of 1,518 regular vetoes and 112 overrides across 1789 through the close of the 116th Congress shows a power held in reserve, exercised rarely, and decisive on the occasions when two thirds of the members present and voting in each chamber chose to exercise the power in the open and by name.

The Threat Does the Work Before Presentment

The formal record of the veto begins at presentment, after both chambers have passed a bill and sent the enrolled text to the president for signature or return. The bargaining record begins much earlier. Long before a measure reaches that formal stage, the possibility of a return with objections shapes the choices made by committees, floor managers, and negotiators in Congress and by officials in the executive branch. The namable claim of the mechanism is that the threat does the work. A president who states an objection in advance, in terms specific enough for drafters to answer, changes the set of provisions that survive drafting, the form in which surviving provisions are written, and the vehicle in which those provisions travel. The returned bill and the override vote, the two steps visible in the constitutional text, then register only the disagreements that advance communication failed to settle. Bargaining before presentment is therefore not an informal addition to the constitutional procedure. It is the predictable consequence of that procedure, once its voting arithmetic, its all or nothing character, and its placement at the end of the legislative sequence are taken together and understood by both branches.

The arithmetic gives the threat its force. Under the rule stated in the constitutional text and interpreted through congressional procedure, tradition, and judicial ruling, an override requires two thirds of the members present and voting, a quorum being present, in the chamber where the bill originated and then two thirds of the members present and voting, a quorum being present, in the other chamber. Failure at either stage ends the effort. The Senate Historical Office table, covering 1789 through the close of the 116th Congress, records 1,518 regular vetoes and 112 overrides across that span. The ratio is about 7.4 percent, roughly one regular veto in fourteen. Pocket vetoes stand outside that calculation because a pocket veto, by definition, involves no return and allows no reconsideration, and the same table separates the 1,066 pocket vetoes in the span from the regular returns. Placed beside the constitutional threshold, the historical ratio describes a burden that falls on Congress after a return. The president defends a decision already made. Supporters of the bill must assemble an extraordinary majority twice, in sequence, under recorded votes, while the president must only hold enough support in one chamber to prevent the second count, or the first, from reaching two thirds.

The sustaining bloc follows from the same arithmetic. If opponents of an override hold one third plus one of the members present and voting in either chamber, the two thirds threshold cannot be met in that chamber, and the bill fails regardless of the margin recorded in the other chamber. The size of that bloc shifts with attendance and with the number of members voting, because the denominator is the membership present and voting rather than the full membership of the chamber, but the structural point holds across those variations. A president assessing a pending bill therefore reads the likely override count rather than the likely passage count. A bill may command a comfortable simple majority in both chambers and yet fall well short of the margin needed after a return. Conversely, a president facing a bill with demonstrated two thirds support in both chambers confronts a threat of limited credibility, because the return would invite a prompt and successful reconsideration. Drafters perform the mirror calculation from the other side of the same table. A provision that draws a stated presidential objection gets weighed against the votes available to defend the provision after a veto, and that weighing begins while the language is open to amendment.

Drafting in the shadow of the threatened objections differs from drafting aimed only at passage. Passage asks whether a majority in each chamber accepts the text. Survival after presentment asks a harder question, whether the text can either win presidential signature or attract two thirds in both chambers after a return. Where neither route appears available, the rational response during drafting is concession in advance, revision of the disputed language, or abandonment of the disputed provision. The concession carries a cost, because the provision presumably entered the bill to serve a purpose its sponsors valued, and narrowing or dropping the provision dilutes that purpose. Refusal to concede carries a different cost, the loss of the entire measure at presentment if the objection proves firm and the override coalition falls short. Between those costs, sponsors choose case by case, and the pattern of choices leaves a trace that the veto statistics alone fail to capture. Bills that embodied compromises with stated executive objections arrived at presentment in signable form and entered the statute books without a veto entry. Bills whose sponsors judged the objection negotiable, overstated, or worth contesting supplied the smaller set of returns and override attempts recorded in the historical table.

The instruments of the threat vary in formality, audience, and timing, and each instrument occupies a distinct place in the sequence before passage and presentment. The most formal instrument is the statement of administration policy, a written communication through which the administration states its position on pending legislation. Such a statement can record support, record opposition, identify specific provisions of concern, propose changes, and state that advisers would recommend a veto if the bill reached the president in a described form. The formulation matters. A recommendation announced by advisers signals a serious objection while preserving a formal step between the statement and a presidential decision, and drafters read that signal against the specificity of the objections attached to the statement. A statement that names a section, describes the defect attributed to the section, and identifies an acceptable alternative gives committees a drafting problem with defined terms. A statement that records general dissatisfaction without particulars gives committees less to answer and leaves more room for disagreement about what concession, if any, would remove the threat.

Public presidential statements form a second instrument, less formal in structure and broader in audience. A president who fixes a position in public, in a speech, a press exchange, or a written message addressed to the public as much as to Congress, converts a negotiating position into a commitment that carries reputational weight. Retreat from a publicly fixed position imposes costs of a different kind from retreat from a staff level communication, because the audience for the reversal extends beyond the negotiators in the room. For that reason, public statements tend to mark the firmer end of the threat spectrum, and congressional readers treat the wording with corresponding care. The conditional phrasing of a public statement deserves the same close reading as the conditional phrasing of a formal policy statement. A declaration tied to a described feature of a bill leaves room for the bill to change and for the position to change with the bill. A declaration tied to the measure as a whole, without qualification, signals that revision of particulars alone may fail to avert a return.

Private negotiation forms the third instrument, and it runs alongside the formal and public channels through much of the drafting sequence. Conversations between executive officials and committee leaders, between the president and congressional leaders, and between staff charged with the text allow objections to be tested before either side commits to a written position. Private negotiation also allows Congress to probe the firmness of a threatened objection. A concern repeated across private exchanges, formal statements, and public declarations presents a different profile from a concern raised once and then left unmentioned as drafting proceeds. The three instruments work best, from the perspective of either branch seeking a signed law, in a legible sequence. Early private exchanges identify the disputed ground. Formal statements record the administration position at defined stages of committee and floor action. Public statements settle the questions that the president chooses to settle beyond recall. Each stage narrows the area of uncertainty for the next, and presentment then tests a text that has already passed through successive filters of stated executive reaction.

Anticipatory concession is the visible congressional response to that sequence, and it takes three recurring forms. A provision may be redrafted, so that the substance survives in altered language designed to remove the stated objection, as where a mandate becomes a reporting requirement, a fixed date becomes a phased schedule, or an open ended authorization becomes a capped and conditioned one. A provision may be dropped, so that the bill proceeds without the contested element and the sponsors reserve the element, if at all, for a separate vehicle or a later Congress. A provision may be narrowed, so that its coverage, duration, or financial scope contracts to a range the stated objection leaves open. Each form appears at the committee stage, where markups rewrite text line by line and where the cost of revision runs lowest, and each form appears again at the floor stage, where amendments offered or accepted in response to a stated objection reshape the text before passage. The timing of a concession affects its price. A change made in committee attracts limited attention and preserves floor time.

The selection effect follows directly. The bills that reach the president form a set already screened by the threat process. Measures that could not pass either chamber never arrive. Measures that passed but embodied objections the president had stated in firm terms arrive only where sponsors either secured two thirds support in advance, accepted the risk of a failed override, valued the recorded confrontation for reasons outside enactment, or misjudged the firmness of the threat. Measures that passed after anticipatory concession arrive in a form the threat helped to write. The population of signed laws therefore contains a large share of texts whose final language reflects executive objections answered during drafting, alongside a share whose sponsors held the votes to disregard those objections. The population of vetoed bills, by contrast, overrepresents miscalculation, deliberate confrontation, and late changes that reopened settled objections after the bargaining sequence had closed. Reading the two populations together explains why the unsigned bill at presentment occupies a distinctive place in the process. An unsigned bill tends to mark either a bargaining failure, in which each side priced the other side incorrectly, or a bargaining refusal, in which at least one side preferred the public record of a veto and an override attempt to the available compromise.

Counting vetoes cast therefore understates the power the threat exercises. The Senate Historical Office figures for 1789 through the close of the 116th Congress count 2,584 vetoes in total and 112 successful overrides, and those counts describe events at the end of the process with precision. The counts carry no entry for the provision withdrawn in committee after a statement of administration policy named the provision as grounds for a recommended veto. The counts carry no entry for the amendment accepted on the floor to narrow a program to a scale the administration had signaled acceptance of in private negotiation. The counts carry no entry for the bill that sponsors declined to schedule at all after a public statement fixed an opposing position and the sustaining bloc took shape in plain view. Each omitted episode represents the veto power operating at full effect, with the return itself rendered unnecessary by the credibility of the threat behind the return. Analysts who measure presidential influence over legislation by veto frequency alone invert the underlying relationship. A low veto count in a given Congress may reflect weakness, agreement between the branches, or a threat record so consistent that drafters internalized the objections before drafting began. The count alone cannot distinguish among those explanations.

Packaging changes the calculation on both sides, because presentment allows no partial action on the enrolled bill. The president signs the measure as passed or returns the measure as passed. A provision the president opposes travels differently alone than inside a large measure carrying provisions the president seeks, and the difference grows with the value the president assigns to the surrounding contents. Standing alone, an opposed provision presents the president with a clean choice, and a return costs the president only the provision itself. Embedded in a measure that funds programs the president supports, authorizes activities the president requested, or carries provisions the president has publicly endorsed, the same opposed provision forces a bundled choice. Signature enacts the objectionable language along with the desired language. Return sacrifices the desired language along with the objectionable language and places the burden of reconstruction on Congress under the pressure of whatever deadlines the measure was written to meet. The mechanics of assembling such combined measures, and the riders that ride within them, receive extended treatment in the discussion of omnibus legislation and the riders it carries, and the bargaining weight of the combined form derives from that assembly process rather than from any feature of the opposed provision itself.

Packaging cuts in both directions, and neutral description requires attention to both edges. For congressional sponsors, packaging raises the price of a veto and can extract signature for a provision that would fail as a freestanding bill. The strategy also carries exposure. A veto of the combined measure returns the entire package, including provisions with broad support, and the override campaign must then defend the package as a whole rather than the contested provision alone. For the president, packaging tests the credibility of earlier threats. A threat stated against a provision in isolation may lose force when the provision arrives inside a measure the president has strong reasons to sign, and drafters who observe a threat discounted in one bundled case adjust the weight assigned to comparable threats afterward. The all or nothing character of presentment is the source of the force on each side. Because no mechanism permits the president to sever the opposed provision from the enrolled text and sign the remainder, the contest shifts from the merits of the single provision to the comparative value each branch assigns to the whole measure, and to the votes available for an override of the whole measure after a return.

The adjacent lever of impoundment requires a sharp distinction, because the lever operates at a different point in the sequence and rests on a different claim of authority. A veto prevents a passed bill from becoming law unless two thirds of both chambers enact the bill over the return. Impoundment, a refusal to spend appropriated funds, operates after enactment, against a law already in force, and concerns execution rather than enactment. The distinction matters for the bargaining analysis as well. A veto threat influences the writing of a bill before the bill binds anyone. An impoundment dispute concerns funds that Congress has already directed through a completed lawmaking process, and the question presented is whether the executive branch may withhold what the enacted text appropriated. Congress answered that question through legislation, in the Congressional Budget and Impoundment Control Act of 1974, the statute examined in the guide to the Congressional Budget Act of 1974. Treating impoundment as a species of veto blurs the line between the qualified negative the Constitution grants at presentment and the separate controversy over execution of enacted appropriations law, a controversy Congress addressed through its own statutory framework in 1974 rather than through the override mechanism.

Congressional counter moves answer the threat within the same arithmetic that gives the threat force. The most direct counter move is construction of the two thirds coalition in advance of passage. Sponsors who assemble override strength during the drafting stage convert the presidential calculation, because a return against a bill with demonstrated two thirds support in both chambers invites a successful reconsideration and the political cost that accompanies a defeated veto. Coalition construction of that kind demands broader agreement than ordinary passage, and the breadth itself shapes the text, since the marginal members needed to move a coalition from a simple majority to two thirds often hold reservations that sponsors answer through the same redrafting and narrowing techniques applied to executive objections. A second counter move is sequencing. Congress controls the order and the packaging of its own measures, within the rules of each chamber, and sponsors may sequence measures to isolate a veto, sending a contested provision in a form that forces the president to accept or return that provision on its own terms rather than inside a bundle of unrelated contents. Isolation reverses the packaging strategy. The veto then stands alone for judgment, and the override campaign addresses a single question instead of a combined measure.

A third counter move attaches an opposed provision to a measure the president has committed to sign. Commitment may arise from a public endorsement, from a request the president transmitted to Congress, or from deadlines embedded in the subject matter of the measure itself. Attachment of that kind confronts the president with the bundled choice described in the packaging analysis, and the strength of the counter move depends on the firmness of the commitment and the cost of the delay that a return would impose. Each counter move has a corresponding weakness that presidents exploit through the instruments of the threat. Advance coalition building consumes time and dilutes text. Isolation surrenders the protection that packaging affords to the rest of the legislative program. Attachment invites a public campaign in which the president separates the attached provision from the host measure in argument and assigns responsibility for the resulting delay to the attachment itself. The bargaining before presentment is therefore symmetrical in structure even where the branches hold unequal formal powers. Congress commands passage, sequencing, and packaging. The president commands signature or return, exercised through a threshold that gives a disciplined minority in one chamber decisive weight after a return.

The threat does the work because the constitutional procedure makes the end of the process visible from its beginning. Every participant in drafting can compute the two thirds requirement, consult the historical record in which roughly one regular veto in fourteen ended in override across the Senate Historical Office table for 1789 through the close of the 116th Congress, and estimate the sustaining bloc available to a president who states an objection early and holds to the statement through passage. Statements of administration policy, public declarations, and private negotiation transmit the objection. Anticipatory concession records the objection inside the text. Packaging and counter packaging redistribute the costs of acting on the objection. Impoundment stands apart from the sequence as a post enactment lever that Congress answered by statute in 1974. The veto entries in the historical table count only the cases in which those transmissions, concessions, and calculations failed to produce a signable bill. The larger part of the institution operates upstream of the count, in committee rooms and floor negotiations where the prospect of a return, rather than the return itself, settles the language that reaches presentment.

The Pocket Veto, Its Litigation, and the Practice That Displaced It

The pocket veto begins in the same sentence of Article I, Section 7 that creates the ordinary return veto, and the difference between the two outcomes turns on a single condition. A bill passed by the House and the Senate is presented to the president before the bill becomes law. If the president approves the bill, the president signs the bill and the measure becomes law. If the president disapproves the bill, the constitutional procedure directs the president to return the bill, together with objections, to the chamber in which the bill originated. A third possibility arises from inaction paired with time. If the president neither signs nor returns the bill within ten days, Sundays excepted, the bill becomes law in the same manner as if the president had signed the bill, unless Congress by adjournment prevents the return of the bill within that period. Where Congress by adjournment prevents return, the bill does not become law. That exception is the pocket branch. There is no signature, there is no return, and there is no statement of objections delivered to the originating chamber for entry on a journal. The measure fails at the end of the ten day period because the condition for automatic enactment has been displaced by the condition for nonenactment, and the constitutional text assigns the decisive role to the adjournment and to its effect on return.

Every element of that outcome contrasts with the elements of the regular veto, and the contrast explains why the two forms carry different consequences for Congress. A regular veto is an act of return. The president sends the bill back to the originating chamber and accompanies the bill with objections that state the reasons for disapproval. That chamber enters the objections at large on its journal, a step that creates a formal legislative record of the disagreement between the branches. The chamber then proceeds to reconsider the bill in light of the objections. If two thirds of the members present and voting in that chamber, a quorum being present, agree to pass the bill, the bill goes together with the objections to the other chamber. The other chamber reconsiders the bill, and approval by two thirds of the members present and voting in that chamber, a quorum being present, enacts the bill notwithstanding the objections. A pocket veto contains none of those steps. Because nothing was returned, there are no objections to enter. Because there are no objections entered on a journal, there is no reconsideration structured by the return procedure. Because there is no reconsideration, there is no override vote in either chamber. The absence of an override follows from the structure of the text. The override procedure presupposes a returned bill, and the pocket branch is defined by the absence of return.

The litigation over the pocket veto therefore centered on a narrow and practical question rather than on the existence of the branch itself. The constitutional text does not grant a free standing power to kill a bill by silence under all circumstances. Silence produces law after ten days, Sundays excepted, in the ordinary case. Silence produces nonenactment only where Congress by adjournment prevents return. Prevention is the operative word, and prevention raises questions of fact and of institutional arrangement. An adjournment might be long or brief. An adjournment might fall between sessions of a Congress or within a session. A chamber might disperse without any provision for receiving papers, or a chamber might leave a designated officer in place with authority to receive messages and returned bills on behalf of the chamber. Read together, the decisions do not treat adjournment as a word with a single automatic effect. The decisions treat adjournment as a condition whose effect depends on whether a return could be made through the arrangements that Congress had provided.

The first of those decisions, The Pocket Veto Case, 279 U.S. 655 (1929), upheld a pocket veto that followed an intersession adjournment. An intersession adjournment is the adjournment that separates one session of a Congress from the next session of the same Congress, and the case required the Supreme Court to decide whether a pocket veto could rest on that kind of adjournment or whether the branch was confined to the final adjournment that ends a Congress. The Court sustained the veto. In doing so, the Court rejected the narrower reading that would have limited the pocket branch to the final adjournment of a Congress. The operative question, under the decision, is whether the adjournment prevents return, not whether the adjournment is the final adjournment of a Congress. That formulation moves attention away from labels and toward function. The label attached to an adjournment, whether intersession, intrasession, or final, does not by itself decide the case. The effect of the adjournment on the ability to return the bill decides the case. At the same time, the decision did not hold that every adjournment prevents return. The decision held that the adjournment before the Court had that effect on the arrangements before the Court, and the qualification concerning arrangements forms part of the holding.

The second part of The Pocket Veto Case concerned delivery to an officer or agent during the adjournment. On the arrangements before the Court, delivery to an officer or agent during that adjournment did not satisfy the return requirement. The point requires careful statement, because a broader statement would misdescribe both the 1929 decision and the decision that followed in 1938. The Court in 1929 assessed the particular intersession adjournment and the particular provision, or lack of provision, for receipt that accompanied that adjournment. Under those conditions, tender to an officer or agent did not complete a return to the originating chamber within the meaning of Article I, Section 7. The reasoning preserved a distinction between delivery to an individual who happens to hold a legislative office and return to the chamber as an institution. If the chamber has adjourned in a manner that prevents that institutional act, physical custody of a document by an officer or agent does not by itself repair the prevention. The 1929 decision thus established two propositions that later cases had to hold at the same time. An adjournment other than a final adjournment may prevent return, and whether an alternative delivery satisfies return depends on the arrangements through which Congress has provided for receipt.

Wright v. United States, 302 U.S. 583 (1938), presented a different arrangement and a different absence, and the Supreme Court reached a different result on those facts. In Wright, a bill was returned to an officer of the originating chamber during a three day Senate recess, and the Court upheld the return. The length of the absence formed part of the context. A recess of three days is a brief interruption within the work of a session rather than the separation between sessions that had been before the Court in 1929. The character of the absence also differed, because the Senate in Wright had provided an officer through whom a return could be received. The Court held that Congress may act through appropriate agencies to receive a return, and that the Constitution does not forbid that practice. Agency, in that reasoning, is not a convenience invented by the executive branch and imposed on Congress. Agency is a means chosen by Congress for the performance of a congressional function. The return was therefore a return to the chamber through the channel the chamber had provided, rather than an attempted delivery to an individual without institutional authority to receive the bill for that purpose.

Wright qualifies The Pocket Veto Case without erasing the earlier decision, and the qualification is best understood through the two variables that separate the cases. The first variable is the length and character of the absence. The 1929 case involved an intersession adjournment, an absence that divides sessions. Wright involved a three day recess, a brief pause that leaves the session intact. The second variable is the existence of a designated recipient. In Wright, an officer of the originating chamber stood in a position to receive the return on behalf of the chamber. In the 1929 case, on the arrangements before the Court, delivery to an officer or agent did not satisfy the requirement. The two decisions together do not produce a rule that an adjournment either always prevents return or never prevents return. The two decisions produce a test that asks whether Congress, through the structure of the adjournment and through the provision made for receipt, left a functioning route by which the president could return the bill to the originating chamber. The pocket branch survived Wright, because Wright did not deny that an adjournment may prevent return. Wright narrowed the circumstances in which prevention would be found by recognizing that Congress may keep the route of return open through an agent even while members are absent.

Kennedy v. Sampson, 511 F.2d 430 (D.C. Cir. 1974), carried the reasoning of Wright into a dispute over an intrasession adjournment. An intrasession adjournment is an adjournment taken within a session, and the question was whether that kind of adjournment prevented return where arrangements existed for receipt of the veto message. The United States Court of Appeals for the District of Columbia Circuit held that an intrasession adjournment does not prevent return where arrangements exist for receipt of the veto message. The holding follows the functional direction set by the earlier cases. If receipt has been arranged, return has not been prevented, and the condition on which the pocket branch depends has not been met. Kennedy v. Sampson is appellate authority rather than a Supreme Court decision, and the distinction in authority matters for an accurate account of the doctrine. The decision is a decision of a federal court of appeals, not a judgment of the Supreme Court, and the decision should be identified at that level whenever the decision is cited. The decision applied the agency principle recognized in Wright to an intrasession setting and treated the existence of receipt arrangements as the fact that defeated a claim of prevention.

A related district court decision, Kennedy v. Jones, 412 F. Supp. 353 (D.D.C. 1976), belongs in the account at a limited and specified strength. In that litigation, challenged pocket veto claims of Presidents Nixon and Ford were invalidated after the Department of Justice agreed to summary judgment. The verified strength for the episode ends there, and a fuller description of reasoning or of remedy would exceed the verified record for the article. Stated at that strength, the 1976 outcome is consistent with the direction of Kennedy v. Sampson two years earlier.

The practice that grew from that litigation is administrative in form and constitutional in effect. Both chambers routinely designate agents authorized to receive messages and returned bills during adjournments and recesses. The agents named in the practice include the Clerk of the House and the Secretary of the Senate. The Clerk serves as the receiving officer for the House, and the Secretary serves as the receiving officer for the Senate. Designation is an act of the chamber concerned, performed in advance of an absence rather than improvised after a bill has been presented. Advance designation matters because the prevention test turns on the arrangements in place during the ten day period. If a chamber has authorized an officer to receive a returned bill, a route of return exists throughout the adjournment or recess, and the president has a known destination for the bill and the objections. The designation does not amend Article I, Section 7, and the designation does not abolish the pocket branch as a textual matter. The designation changes the factual predicate on which the branch depends. Because receipt is arranged, adjournment rarely prevents return.

That change in factual predicate explains the transformation in the character of pocket veto disputes after the development of the agent for receipt practice. The undisputed silent pocket veto has largely disappeared. A silent pocket veto, in the classic form, involved no return, no objections, and no congressional proceeding on the bill after the expiration of the ten day period. The bill simply did not become law, and Congress had no returned measure to reconsider. Such an outcome depends on prevention that no participant disputes on the facts of receipt. Where both chambers maintain designated recipients during absences, prevention is difficult to assert for an ordinary adjournment and difficult to sustain against the arrangements that Congress has placed on the record. The disappearance described here is a disappearance of the undisputed form, not a claim that no president after a particular year attempted to invoke the pocket branch. Attempts continued in later administrations, and the disputed episodes show a different pattern.

The disputed claims associated with President George H. W. Bush illustrate the first part of that later pattern. Per the Congressional Research Service, President George H. W. Bush attempted two pocket vetoes during intrasession recesses. Congress considered the bills enacted because the bills had not been returned. The reasoning attributed to Congress in that account follows directly from the constitutional default. If return was not prevented, and if the president neither signed nor returned a bill within ten days, Sundays excepted, the bill becomes law as if signed. An intrasession recess with receipt arrangements in place does not supply prevention under the reasoning of Wright and Kennedy v. Sampson, and a claim of pocket veto under those conditions competes with the automatic enactment rule rather than with the override rule. The disputed claims are not counted in the Congressional Research Service historical table, a treatment that reflects the disputed status of the claims rather than an agreed classification as completed pocket vetoes. The episode therefore shows prevention contested at the threshold. Acceptance of prevention would have meant nonenactment without return. Rejection of prevention, combined with the absence of return, meant enactment by expiration of the period.

The episode of December 28, 2007 presented the pattern in a more elaborate form, the protective return. On that date President George W. Bush issued a memorandum of disapproval for H.R. 1585, the fiscal 2008 defense authorization bill. The memorandum claimed that adjournment had prevented return and invoked The Pocket Veto Case in support of the claim. At the same time, the president returned the bill with objections to the Clerk of the House. The stated purposes of the simultaneous return, as recorded in the verified account, were to leave no doubt that the bill was vetoed and to avoid litigation. The pairing is the defining feature of the episode. The December 28, 2007 action combined both positions in a single transaction, asserting prevention while performing the act that prevention would have made impossible. Congress treated H.R. 1585 as a regular veto, and the Congressional Research Service treated H.R. 1585 as a regular veto. That treatment followed the performed act. A bill returned with objections to the Clerk of the House, an agent designated to receive returned bills for the House, presents the originating chamber with the material required for journal entry and reconsideration. Classification as a regular veto preserved the override procedure that a pocket classification would have foreclosed.

The episode of October 8, 2010 followed a related course without the same memorandum structure. On that date President Barack Obama characterized the veto of H.R. 3808 as a pocket veto. The House treated the action as a regular veto and held an override vote on November 17, 2010, and the override vote failed. The contrast between characterization and treatment is the central fact of the episode. Characterization by the president asserted the pocket branch. Treatment by the House applied the regular veto framework, under which a returned bill is reconsidered and may be enacted over objections by two thirds of the members present and voting in each chamber, a quorum being present. The episode nevertheless demonstrates the institutional consequence of a return accompanied by objections and received through the channels of the originating chamber. Once the House had the bill and the objections in a form that permitted reconsideration, the House proceeded under the reconsideration procedure. As in 2007, the practical classification followed the return rather than the label attached to the return.

The protective return pattern shared by the disputed modern episodes can be stated without assigning motive beyond the purposes recorded for the 2007 action. In a protective return, a president who asserts that adjournment has prevented return also returns the bill and the objections to the designated officer of the originating chamber. The return protects the veto against the possibility that a court or Congress will reject the claim of prevention. If prevention is accepted, the bill has not become law under the pocket branch. If prevention is rejected, the timely return with objections stands as a regular veto, and the bill has not become law by expiration of the ten day period. Without the protective return, rejection of prevention would produce the opposite result, because a bill neither signed nor returned within ten days, Sundays excepted, becomes law as if signed. The protective return therefore separates disapproval from classification. Congress and the Congressional Research Service, in the 2007 episode, resolved classification in favor of the regular veto. The House, in the 2010 episode, proceeded on the regular veto classification and voted. The George H. W. Bush episodes show the risk that the protective return addresses, because claims made without an accepted return led Congress to treat the bills as enacted.

The doctrine and the practice together leave the pocket veto in a position that is narrower than the position suggested by the text read in isolation and broader than a claim of abolition would suggest. The text retains the branch. Where Congress by adjournment prevents return within ten days, Sundays excepted, a bill does not become law, and no override is possible because nothing was returned. The Pocket Veto Case, 279 U.S. 655 (1929), retains the proposition that an intersession adjournment may supply that prevention and that the test is prevention rather than final adjournment. Wright v. United States, 302 U.S. 583 (1938), retains the countervailing proposition that Congress may act through appropriate agencies to receive a return, including receipt by an officer during a three day recess of the originating chamber. Kennedy v. Sampson, 511 F.2d 430 (D.C. Cir. 1974), as appellate authority, extends the receipt principle to an intrasession adjournment where arrangements exist for receipt of the veto message, and Kennedy v. Jones in 1976 records invalidation of challenged claims after agreement to summary judgment. Because receipt is arranged in the ordinary case, adjournment rarely prevents return, the undisputed silent pocket veto has largely disappeared, and disputed claims from the administrations of President George H. W. Bush, President George W. Bush, and President Barack Obama were answered through enactment treatment, regular veto treatment, or an override vote rather than through acceptance of silent nonenactment.

How did designated agents for receipt change pocket veto disputes?

Both chambers began designating agents in advance to receive messages and returned bills during absences, including the Clerk of the House and the Secretary of the Senate. With a recipient in place, return was harder to call prevented, and the undisputed silent pocket veto largely disappeared from practice.

The Line Item Veto Experiment and Its Invalidation

The demand that produced the federal line item veto experiment arose from the structure of presentment itself. Under Article I, Section 7 of the Constitution, every bill passed by the House of Representatives and the Senate is presented to the president before the bill becomes law. Signature makes the measure law. Return with objections sends the measure back to the chamber where the measure originated, where the objections are entered at large on the journal and the measure is reconsidered. Agreement by two thirds of the members present and voting in the originating chamber, followed by approval by two thirds of the members present and voting in the other chamber, enacts the measure over the objections. The choice attaches to the measure as a whole. Presentment offers the whole measure or nothing. No constitutional provision grants the president authority to sign part of a measure into law and reject another part, to strike a single appropriation while approving the balance of an appropriations act, or to remove a discrete tax provision while leaving the rest of a revenue measure intact.

That whole measure structure shaped the argument for a different authority. Large measures, and appropriations measures in particular, combine many separable decisions into a single text presented for a single decision. A president who objected to one dollar amount, one project, or one narrow benefit faced a disproportionate alternative under the ordinary procedure. Rejection of the entire measure would defeat every other provision in the measure, including provisions the president supported and provisions carrying funding for ongoing government operations. Supporters of item veto authority described that structure as a gap between the scale of congressional packaging and the scale of the presidential negative. The proposed remedy was authority to reject individual items within a signed measure, so that objection could be directed at the item rather than at the measure as a whole. Presidents of both parties had sought such authority, and the request therefore appeared in the historical record as an institutional request by occupants of the presidency rather than as a position belonging to one party alone. The argument was stated in terms of fiscal control and accountability.

A contrast with state practice supplied another part of the demand, and the contrast requires precision. Most state governors hold some form of item veto under state constitutions. Many governors confronting a state appropriations measure possess a constitutionally grounded power to disapprove items within the measure while approving the remainder. The federal president holds no equivalent power under the United States Constitution. The difference does not result from silence that a statute might fill without difficulty. The federal procedure is set out in the Presentment Clause, and the procedure specifies passage by both chambers, presentment of the passed bill, and then signature, return, enactment without signature under the ten day rule, or enactment over a return veto by the prescribed two thirds votes. A state governor exercises an item veto because a state constitution confers that power. The Line Item Veto Act of 1996 represented the statutory attempt, and the litigation that followed tested whether a statute could supply what the constitutional text had not supplied directly.

Congress enacted the Line Item Veto Act of 1996 as Public Law 104-130, cited at 110 Stat. 1200. The Act was signed on April 9, 1996, and took effect on January 1, 1997. The Act was codified in part at 2 U.S.C. sections 691 to 692. The measure was an ordinary statute, enacted through the ordinary lawmaking process and signed by the president. The authority created by the statute therefore originated in legislation, not in a constitutional amendment. The delayed effective date placed the start of the authority at the beginning of 1997, after the 1996 enactment and signature. The statute announced a new presidential authority, defined the categories subject to the authority, imposed determinations that had to precede exercise of the authority, created a message procedure for communicating an exercise to Congress, provided for effectiveness on receipt of the message, and established a congressional response in the form of a disapproval bill. The constitutional question that reached the Supreme Court in 1998 concerned that entire arrangement, and in particular the legal effect of a presidential cancellation issued after a bill had been signed and had become law.

The cancellation authority itself was defined by category, and the categories were narrower than a general power to strike any provision from any law. The president was authorized, within five calendar days after signing a law, to cancel in whole one of three things. The first category was a dollar amount of discretionary budget authority. The second category was an item of new direct spending. The third category was a limited tax benefit. The verified record describes the authority as cancellation in whole. The five calendar day window also defined the authority in time. Cancellation was not an open ended power to revisit laws at any point during a presidency. The window ran after signing, a sequence that proved central to the constitutional analysis. Under the Act, the president first signed the bill, so the bill became law under the ordinary presentment procedure. Only after that completed act of enactment did the cancellation window open. A decision made before signature to return a bill with objections followed the constitutional return veto path. A decision made after signature to cancel an item followed the statutory path created in 1996, and the difference between a negative exercised before a measure becomes law and a cancellation exercised after a measure becomes law supplied the distinction on which the experiment ultimately foundered.

Before canceling an item, the president had to make three determinations, and the determinations functioned as statutory conditions on the authority. The president had to determine that the cancellation would reduce the federal deficit. The president had to determine that the cancellation would not impair essential government functions. The president had to determine that the cancellation would not harm the national interest. The three determinations addressed three different kinds of consequence. Deficit reduction tied the authority to a fiscal purpose and presented the cancellation as a measure directed at the size of the federal deficit rather than at policy disagreement alone. Protection of the national interest imposed the broadest of the three conditions, requiring a determination of no harm to a general interest extending beyond budget arithmetic and beyond the operation of a particular program. The determinations were presidential findings required by the statute as a predicate to cancellation. Once the determinations were made and the cancellation was communicated in the manner the statute prescribed, the cancellation took effect under the statutory mechanism, subject to the congressional response that the statute also prescribed.

Communication took the form of a special message to Congress. A cancellation took effect on receipt by Congress of the special message. The receipt rule matters because effectiveness did not depend on enactment of a further law approving the cancellation. The sequence ran in the opposite direction from ordinary legislation. In ordinary legislation, a text has no legal effect as a statute until the text has passed both chambers and has completed presentment by signature, by expiration of the ten day period without adjournment preventing return, or by override of a veto. Under the Line Item Veto Act, by contrast, the underlying bill had already completed that process and had become law at signature. The special message then produced a change in the law that was already in effect, by canceling an amount, item, or benefit that Congress had enacted and the president had signed. Effectiveness on receipt placed the operative moment at delivery of the presidential message to Congress. That design was also the source of the constitutional defect identified in 1998, because the legal change produced by the message altered the content of enacted law without passage of a new bill through the procedure prescribed for amendment or repeal.

Congress retained a route back under the statute, and that route was the disapproval bill. Congress could pass a disapproval bill restoring the canceled item through the ordinary process. A disapproval bill was a bill. Passage by both chambers and presentment to the president were therefore required, and the president could veto the disapproval bill. If the president vetoed a disapproval bill, two thirds of each chamber would then be needed to override the veto and enact the restoration over presidential objection. Congress enacted a measure containing an item. The president signed the measure and then, within the statutory window and after making the required determinations, canceled the item by special message, with the cancellation effective on receipt. Congress responded by passing a bill disapproving the cancellation and restoring the item. The president, who had exercised the cancellation, could return the disapproval bill with objections. Congress could restore the item only by completing the override procedure in both chambers. The burden of assembling supermajorities therefore shifted with the stage of the sequence. An item enacted by ordinary majorities and signed into law could be canceled by presidential action under the statute, and restoration of the item required either presidential signature of the disapproval bill or two thirds votes in both chambers over a veto. The Supreme Court analysis in 1998 did not turn on a policy preference between those descriptions. The analysis turned on whether the Constitution permits a statute to authorize that sequence at all.

Use of the Act before invalidation supplied a concrete record against which the mechanism can be described without speculation. President Clinton sent 11 special messages canceling 82 items in total, according to the Congressional Research Service figures carried in the verified record. The numbers establish scale and form. Cancellation was exercised repeatedly, through multiple messages, and the unit of counting was the item canceled rather than the message sent. The significance of the figures for the procedural history is narrower and firmer. The authority was not enacted and left unused. The authority was exercised, cancellations took effect under the statutory design, and Congress used the statutory response in at least one instance carried through to enacted law. The experiment therefore reached the courts and the congressional record as an operating mechanism, not as a hypothetical procedure described only in statutory text.

The single enacted disapproval law illustrates the disapproval mechanism at full length. Public Law 105-159 disapproved cancellation of 38 military construction projects. The law was enacted over the president’s veto. The override votes were 347 to 69 in the House of Representatives and 78 to 20 in the Senate, according to the Congressional Research Service figures carried in the verified record. Three different actions are compressed in that summary. First, items concerning 38 military construction projects had been canceled under the Line Item Veto Act. Second, Congress passed a disapproval bill directed at those cancellations, seeking restoration through legislation. Third, the president vetoed the disapproval bill, and Congress overrode the veto by margins exceeding two thirds in each chamber, so the disapproval bill became Public Law 105-159 despite the veto. Precisely what the episode shows is that the disapproval mechanism ran through ordinary presentment, including an override. Restoration was not accomplished by resolution, by objection, or by the mere introduction of a disapproval measure. Restoration required a bill that passed both chambers, went to the president, was returned with objections, and then received the constitutionally prescribed supermajority support in both chambers. The recorded margins exceeded the two thirds threshold by a substantial amount, with 347 votes for override against 69 in the House and 78 votes for override against 20 in the Senate.

Litigation reached the Supreme Court in two stages, and the first stage ended without a decision on the constitutionality of cancellation. In Raines v. Byrd, 521 U.S. 811, decided in 1997, six members of Congress who had voted against the Act sued to challenge the statute. The Supreme Court held that the challengers lacked standing. The injury asserted by the members was an abstract institutional injury from dilution of legislative power, and the Court held that such an injury was not a concrete personal injury sufficient to support standing. The distinction between institutional injury and personal injury controlled the outcome. The Court treated that claim as abstract and institutional rather than concrete and personal, and the case ended on that threshold ground. The Act survived the first challenge on standing grounds only. Survival on standing grounds carried no holding that cancellation complied with the Presentment Clause, no holding that a signed law could be altered by special message, and no holding on amendment or repeal by presidential action. The merits waited because the party before the Court had not presented a claim the Court would hear.

That posture arrived in Clinton v. City of New York, 524 U.S. 417, decided on June 25, 1998. The Supreme Court decided the case by a vote of six to three, in an opinion by Justice Stevens. The Court held that the cancellation procedures of the Line Item Veto Act violated the Presentment Clause. The reasoning began with the legal status of a bill after signature. Once the president had signed the bills at issue, the bills were law. Signature completed presentment and enacted the text that both chambers had passed and that had been presented to the president. A cancellation issued after that point therefore operated on enacted law. In legal effect, the cancellations amended or partly repealed enacted laws, by removing amounts, items, or benefits that the enacted text contained. The form of the presidential action was a cancellation under a statute and communication by special message. The legal effect, as the Court analyzed the effect, was a change in the funding authorities and other legal consequences established by laws already enacted. Characterization by effect rather than by label placed the cancellation within the constitutional rules governing amendment and repeal, rather than within a separate category of executive action exempt from those rules.

From that premise, the Court applied the lawmaking procedure prescribed by the Constitution. Amendment and repeal of a statute must follow the single, finely wrought and exhaustively considered procedure for making law set out in the Constitution. That procedure requires passage of a bill by both the House of Representatives and the Senate and presentment of the bill to the president, followed by signature, enactment without signature under the conditions stated in the text, or enactment over a veto by two thirds of the members present and voting in each chamber. The procedure is singular in the sense that the Constitution provides one route for enacting, amending, or repealing a statute, and the procedure is exhaustive in the sense that compliance with a different route devised by statute does not substitute for the prescribed steps. Congress could not, by passing the Line Item Veto Act through the correct procedure in 1996, authorize a later change in enacted law through a procedure omitting passage and presentment of the change itself. Each cancellation changed law, and each change, under the Court’s reasoning, required the same lawmaking procedure as any other amendment or partial repeal. A special message effective on receipt did not satisfy that requirement, even though the statute authorizing the message had been enacted in full compliance with the requirement and even though Congress had provided a disapproval bill procedure under which a cancellation could be reversed by a further law.

The holding was stated in direct terms. The president may not amend or repeal a statute unilaterally. Unilateral action in that sentence describes the decisive feature of cancellation under the Act. The president acted alone at the operative moment, after signature, on the basis of statutory authorization, required determinations, and a special message. No chamber passed the cancellation. Congressional participation came before cancellation, in enactment of the authorizing statute, and after cancellation, in the opportunity to pass a disapproval bill subject to veto. The Court held that participation before and after did not cure the absence of the prescribed procedure for the change itself. The six to three division in the decision records disagreement within the Court, but the majority holding determined the legal status of the Act. After June 25, 1998, the cancellation authority could not be exercised as the statute had provided, because the procedures authorizing cancellation violated the Presentment Clause. The decision addressed the federal Constitution and the federal lawmaking process. The decision denied only the proposition that a president, acting under a statute, could produce that change alone after signing the law containing the item.

The consequence can be stated plainly. No federal line item veto exists. The federal president holds no authority to cancel a dollar amount of discretionary budget authority, an item of new direct spending, or a limited tax benefit from a signed law by special message effective on receipt. A statute on the same model as the Line Item Veto Act of 1996 cannot restore such an authority, because Clinton v. City of New York held the model unconstitutional under the Presentment Clause. The route to a federal item veto, if such a power is to exist in a form comparable to the authority exercised between January 1, 1997, and the 1998 decision, is constitutional amendment. An amendment could confer item veto authority directly in the constitutional text, define the categories subject to the authority, and prescribe the effect of a disapproval or override. Absent such an amendment, disputes over individual items in a measure must be resolved through the procedures the Constitution and ordinary legislation provide, by negotiation before passage, by enactment or defeat of the measure as presented, by return veto and possible override, or by a later enacted law amending or repealing the item through full passage and presentment.

The distinction between cancellation after signature and the return veto before signature marks the constitutional boundary preserved by the 1998 decision. The return veto is placed by the Constitution before a bill becomes law. A president who disapproves a bill returns the bill with objections to the originating chamber, the objections are entered on the journal, and Congress decides by the prescribed two thirds votes whether the bill becomes law despite the objections. The negative operates on a proposal that has passed both chambers but has not completed enactment. The president acts on the whole bill, and Congress retains the final decision through override. Cancellation under the 1996 Act operated at the opposite point in the sequence. The president signed the bill, the bill became law, and the president then removed items from the law that signature had enacted. The negative operated on law in effect, and congressional restoration required a new bill subject to another veto. Placement in the sequence therefore determined the constitutional character of the action. A whole measure negative before enactment belongs to the qualified negative the Constitution grants. A selective negative after enactment, exercised by message and effective on receipt, belongs to amendment or partial repeal under the Court’s analysis, and amendment or repeal belongs to the full lawmaking procedure. Presidents retained every power conferred by the Presentment Clause, including signature and return with objections. What presidents lacked after June 25, 1998, and what no statute on the 1996 model could confer, was authority to sign a measure into law and then strike items from the enacted measure alone.

Overrides in the Record, Signing Statements, and Common Misunderstandings

The record of overrides gives the qualified negative a concrete shape that the constitutional text alone does not supply. Article I, Section 7 describes a sequence. A bill passed by the House and the Senate is presented to the president. The president signs the bill, or returns the bill with objections to the chamber where the bill originated. That chamber enters the objections at large on its journal and proceeds to reconsider the bill. Agreement of two thirds of that chamber passes the bill, and the bill goes with the objections to the other chamber. Reconsideration follows there, and approval by two thirds of that chamber enacts the bill. In all override cases the votes of both chambers are determined by yeas and nays, and the names of the persons voting for and against the bill are entered on the journal of each chamber respectively. The recorded vote requirement is written into the constitutional text itself, and the requirement matters for reading the record. An override tally is not an estimate, a voice vote, or a report of sentiment. The margins in the examples that follow therefore carry a particular kind of weight. Each margin records members who voted, on the record, to enact a measure after the president had returned that measure with objections, and each margin had to be assembled twice, once in the chamber of origin and once in the second chamber, under the same two thirds rule.

The first successful override in the record occurred on March 3, 1845, over a veto by President John Tyler. The measure was S. 66, an appropriations measure barring construction of revenue cutters without an appropriation. The House voted 126 to 31 on the final day of the 28th Congress. The override was completed at the close of a Congress, under the ordinary sequence of return, journal entry, reconsideration, and recorded vote, and the House margin crossed the two thirds threshold of members present and voting. A single override in 1845 established that the reconsideration procedure could produce a law over a presidential return.

The first of the three examples in the series is the Freedom of Information Act amendments of 1974, overridden over the veto of President Gerald Ford. The House voted 371 to 31. The Senate voted 65 to 27. The description stops there, with subject and margin, because the assignment for the three examples permits no characterization beyond subject and margin, and the discipline has a reason. A margin records how many members present and voting supported enactment on reconsideration and how many opposed enactment on reconsideration. The margin does not, by itself, describe the merits of a measure, the motives of the members who voted, or the effect of the enacted law. The same restraint governs the two examples that follow, and the restraint keeps the three examples comparable. Each example presents a subject, a president who returned the bill, and two recorded margins. Comparison across the examples is comparison of procedure and arithmetic, and procedure and arithmetic are the subjects of the article.

The second example is the Civil Rights Restoration Act. The override was completed on March 22, 1988, over the veto of President Ronald Reagan. The Senate voted 73 to 24. The House voted 292 to 133. The act responded to Grove City College v. Bell and restored institution-wide coverage under the federal civil rights statutes that the act amended. The order of the two margins in the record follows the path of reconsideration through the two chambers, and the two margins together complete the constitutional requirement. Neither margin alone enacted the law. Enactment required two thirds of members present and voting in the Senate and two thirds of members present and voting in the House, on the same returned bill, with the presidential objections before both chambers. The date of completion, March 22, 1988, marks the point at which the second of the two required margins was recorded and the bill became law through reconsideration rather than through signature.

The third example is the Cable Television Consumer Protection and Competition Act of 1992, S. 12, Public Law 102-385. President George H. W. Bush vetoed the act on October 3, 1992. The override was completed on October 5, 1992. The Senate voted 74 to 25. The House voted 308 to 114. The override was that president’s only override. The interval between October 3, 1992 and October 5, 1992 illustrates a feature of the return procedure that a signature does not share. A returned bill arrives in the originating chamber with objections entered on the journal, and reconsideration is a defined parliamentary step rather than the start of a new legislative effort. The two chambers reconsider the identical bill in light of the objections, and the recorded margins decide whether the bill becomes law. In the 1992 example the two margins were recorded within two days of the veto, and the sequence closed with enactment through the second margin.

One structural observation is permitted by the brief, and the three examples support that observation and no broader claim. An override assembles two thirds of members present and voting in both chambers, which ordinarily requires support beyond the president’s opponents. The arithmetic explains the observation. A president ordinarily retains support among members of the president’s own party and among other members who agreed with the objections, agreed with the president on the subject, or declined to enact a law over a return for institutional reasons. A coalition limited to members who opposed the president on the measure in general terms does not ordinarily reach two thirds in either chamber, and the requirement must be met in two chambers with different memberships, different terms, and different constituencies. The coalition that produces an override therefore ordinarily includes members who had supported the president on other questions, members who had passed the bill originally by a smaller margin and then held firm on reconsideration in larger numbers, or members who changed position after the objections were stated. The record does not assign motives to any of those members in the three examples, and no motive is assigned here. The observation is structural. Two successive supermajority votes, each recorded by name, each taken on a bill that the president had formally opposed in writing, ordinarily demand a breadth of support that opposition alone does not supply.

Signing statements belong to a different part of the process, after a president has chosen signature rather than return. A signing statement is a presidential statement issued on signing a bill. A signing statement sometimes states constitutional objections to particular provisions of the signed law. A signing statement sometimes directs how the executive branch will construe particular provisions of the signed law. A signing statement is not a veto, because the bill becomes law in full. No provision is returned to Congress. No chamber enters objections on a journal in response to a signing statement, no reconsideration follows, and no supermajority vote is required or held. The entire measure, including any provision identified in the statement, enters into law through the signature. That single distinction controls every other question about signing statements, and confusion on the distinction accounts for a large share of misunderstanding about executive power after presentment. A veto prevents enactment unless both chambers override the veto. A signing statement prevents nothing, deletes nothing, and suspends nothing by its own force.

A signing statement has no formal legal effect on the validity of a statute. The statute as enacted through bicameral passage and presentment carries the legal force, and the statement does not amend the text, repeal a provision, or alter the validity of any provision. The interpretive weight of a signing statement is contested. One line of discussion treats a statement as part of the record surrounding enactment and as evidence of executive understanding that courts, agencies, and Congress may consider or may discount. Another line of discussion treats a statement as an executive document issued after passage, without participation by the two chambers that passed the text, and therefore as a weak or improper guide to the meaning of the text. The contest has not been resolved by a single governing rule stated in the fact record for the article, and the article offers no resolution. Presidents of both parties have issued signing statements, and the practice cannot be described accurately as the practice of one party, one administration type, or one period alone.

A 1993 opinion of the Office of Legal Counsel concluded that presidential signing statements serve legitimate and defensible purposes. The opinion included, among the purposes identified, informing Congress that the executive branch considers a provision unconstitutional in certain applications. Notice to Congress forms part of that position. A president who signs a bill containing a provision that the executive branch regards as unconstitutional in certain applications may use a signing statement to place the objection on the record at the moment of signing, to state the construction that the executive branch will adopt, and to give Congress an opportunity to respond through amendment, oversight, or further legislation. Bills presented to the president often contain many provisions on many subjects. Return of an entire bill over one provision carries consequences for every other provision in the measure, including provisions that the president supports and that both chambers passed. A signing statement, on that view, permits signature of the whole measure while preserving a constitutional objection in a public form, without claiming that the objection removes the provision from the law.

In 2006 an American Bar Association task force reported against the practice of using signing statements to claim authority to disregard or decline to enforce provisions of a signed law. The task force described that use as contrary to the rule of law and to the constitutional system of separation of powers. The position addressed a specific claimed use, the claim of authority to disregard or decline to enforce a provision after signing the law that contains the provision, rather than every statement issued on signing. Critics associated with the position argued that a president who regards a provision as unconstitutional faces constitutional choices at presentment, including return of the bill with objections, and that signature followed by nonenforcement converts a qualified negative, which Congress can override, into a selective negative that Congress cannot answer through the reconsideration procedure. Defenders of signing statements answered that enforcement involves construction in concrete applications, that a provision may be valid in some applications and invalid in others, and that a statement of construction at signing gives notice rather than concealment. The two positions, the 1993 Office of Legal Counsel position and the 2006 American Bar Association task force position, are presented here with equal care and without resolution, because the positions turn on contested questions about presentment, enforcement discretion, and the weight of executive interpretation. The factual core shared by both positions is narrow and firm. A signing statement accompanies signature. The bill becomes law in full. The statement has no formal legal effect on validity. Every broader conclusion drawn from that core belongs to one side of a continuing argument or to the other.

The common misunderstandings about the veto and the override repeat a small number of errors in different forms, and each error can be stated and corrected in prose without enlarging the record. The first misunderstanding holds that two thirds means two thirds of all members of a chamber. The Constitution speaks of two thirds of that House, and congressional procedure, tradition, and judicial ruling interpret the phrase as two thirds of members present and voting, a quorum being present. In Missouri Pacific Railway Co. v. Kansas, 248 U.S. 276 (1919), a unanimous Supreme Court, in an opinion by Chief Justice White, held that two thirds of a quorum present suffices for an override, in a case that challenged enactment of the Webb-Kenyon Act over a veto. The practical consequence is that absences and abstentions change the denominator. A chamber with members absent requires fewer affirmative votes to reach two thirds of members present and voting than a chamber at full attendance, and a member who is present but does not vote does not enter the denominator under the stated rule. Every margin quoted in the article, 126 to 31, 371 to 31, 65 to 27, 73 to 24, 292 to 133, 74 to 25, and 308 to 114, is therefore read against attendance on the day of the vote, not against the full elected membership of the chamber. A report that converts an override margin into a fraction of all members without accounting for attendance misstates the constitutional test.

The second misunderstanding holds that a pocket veto can be overridden. A pocket veto arises on the second branch of the ten day rule. If the president neither signs nor returns a bill within ten days, Sundays excepted, the bill becomes law as if signed, unless Congress by adjournment prevents return of the bill, in which case the bill is not law. On that branch nothing was returned, so there is nothing to reconsider. Reconsideration under Article I, Section 7 begins with a returned bill and objections entered on the journal of the originating chamber. A pocket veto produces neither a returned bill nor objections in that form, and the override procedure has no object on which to operate. Congress may pass the same measure again as a new bill in a later session or a later Congress, and the new bill then goes through passage and presentment in the ordinary way. Passage of a new bill is legislation, not an override, and the distinction explains why pocket vetoes are excluded from the denominator in any honest override ratio. A ratio that divides overrides by all vetoes, regular and pocket combined, understates the override rate for returned bills and mixes two procedures that the Constitution treats differently.

The third misunderstanding holds that the president holds a line item veto. The federal president holds no line item veto. Congress attempted to confer a form of item cancellation authority through the Line Item Veto Act of 1996, Public Law 104-130, signed on April 9, 1996 and effective on January 1, 1997. The statute authorized the president, within five calendar days after signing a law, to cancel in whole a dollar amount of discretionary budget authority, an item of new direct spending, or a limited tax benefit, after determining that the cancellation would reduce the federal deficit, would not impair essential government functions, and would not harm the national interest. President Bill Clinton sent 11 special messages cancelling 82 items in total. In Clinton v. City of New York, 524 U.S. 417 (1998), decided on June 25, 1998 by a vote of 6 to 3, in an opinion by Justice Stevens, the Supreme Court held the cancellation procedures unconstitutional under the Presentment Clause. Once signed, the bills were law, and the cancellations in legal effect amended or partly repealed enacted laws. Amendment and repeal must follow the single, finely wrought and exhaustively considered procedure for lawmaking that the Constitution prescribes, and the president may not amend or repeal a statute unilaterally. A statute alone cannot confer a federal line item veto after the 1998 decision. A constitutional amendment would be required.

The fourth misunderstanding holds that a signing statement cancels provisions of a signed law. A signing statement changes no law. The correction follows directly from the definition given in the discussion of signing statements. Signature enacts the bill in full, and a statement issued on signing has no formal legal effect on the validity of any provision. A statement may record a constitutional objection. A statement may direct executive construction. Neither act removes text from the statute, suspends a provision by force of the statement alone, or converts the signed law into a shorter law containing only the provisions that the statement approved. The contrast with the line item cancellation struck down in 1998 sharpens the correction. Cancellation under the 1996 statute purported to eliminate items after signature, and the Supreme Court held that elimination unconstitutional. A signing statement does not even purport, as a formal matter, to eliminate an item, and any enforcement dispute that follows a statement is a dispute about construction and enforcement in a concrete setting, not the product of a textual deletion performed by the statement. A reader who treats an objection recorded in a signing statement as a deletion misreads both the statement and the statute.

The fifth misunderstanding holds that a veto count measures the power of the veto. A count of vetoes cast counts only the returns that occurred. The threat of a veto shapes bills before presentment, so the count understates the power. The mechanism operates through stated administration positions. An administration states a position on pending legislation through statements of administration policy, which can state that advisers would recommend a veto. A public veto threat fixes a bargaining position before passage. Because an override requires two thirds of members present and voting in both chambers, drafters who cannot reach that margin negotiate against the threatened objections before presentment. Provisions are rewritten, narrowed, delayed, or removed during drafting and negotiation, and the enacted bill that reaches the president differs from the bill that would have reached the president without the threat. None of that influence appears in a veto total. The most consequential vetoes, in the phrasing of the verified record, are the ones never cast because the bill changed first. Packaging affects the same leverage. A provision that a president opposes travels differently alone than inside a large measure carrying provisions that the president seeks, and a threat directed at a large measure carries a different cost from a threat directed at a short measure on a single subject. A veto count is a useful record of formal returns, of pocket vetoes where adjournment prevented return, and of overrides completed. The count is not a measure of influence, and influence exercised before presentment leaves no entry in the count at all.

The sixth misunderstanding holds that law without signature signals disapproval. Inaction enacts a bill where Congress can receive a return, whatever the president’s view of the bill. If the president neither signs nor returns a bill within ten days, Sundays excepted, the bill becomes law as if signed, unless Congress by adjournment prevents return. The first branch of the rule contains no inquiry into motive. A president may allow a bill to become law without signature because of opposition to the bill combined with a judgment that a veto would be overridden, because of indifference to the subject, because of support for the bill combined with a wish to avoid a signing act, or for reasons that the record does not disclose. The constitutional result is identical in each case. The bill becomes law in full, with the same force as a signed bill, and the absence of a signature carries no formal legal meaning about approval or disapproval. The pocket veto branch is the exception that defines the boundary. Where adjournment prevents return, inaction defeats the bill. Where return is available through the arrangements that Congress maintains for receipt of messages and returned bills, inaction enacts the bill. A reader who infers disapproval from the absence of a signature alone adds a signal that the procedure does not send, and a reader who infers approval from enactment without signature makes the mirror error. The procedure records an act, signature, return, or the passage of ten days under stated conditions. Meaning beyond the act belongs to evidence outside the procedure, in statements, messages, and the surrounding record, judged on the strength of that evidence rather than on the form of enactment alone.

Reading a Presented Measure: The Questions the Mechanism Asks

The veto and the override reward a reader who treats a presented measure as a procedural object before treating the measure as a political event. Every element of the mechanism examined in the preceding sections points toward the same discipline. The constitutional text fixes a sequence and a clock. The override threshold converts that sequence into arithmetic that drafters, negotiators, and presidents all compute in advance. The pocket veto branch makes the calendar of Congress itself a substantive fact, capable of deciding a measure without a vote or a message. The line item veto litigation marks the boundary of the entire arrangement, because the Supreme Court held in Clinton v. City of New York that amendment and repeal of enacted law must follow the single, finely wrought and exhaustively considered procedure that Section 7 prescribes, and that no statute could hand a president a cancellation power that procedure withholds. Signing statements occupy the space outside the mechanism altogether, carrying no formal legal effect on the validity of a signed law, however contested the interpretive weight of a statement has been between the executive branch position recorded in a 1993 opinion of the Office of Legal Counsel and the contrary position that an American Bar Association task force stated in 2006. A reader who keeps those boundaries separate can place any headline about a veto in the correct category on first contact.

Integration of the bargaining account with the textual account resolves the apparent tension between a power described as rarely exercised and a power described as constantly influential. The rarity of the override and the influence of the threat are products of the same threshold. Because two thirds of the members present and voting in both chambers is a demanding margin, a stated objection reorganizes drafting incentives far upstream of presentment, and measures arrive at the presidential desk preadjusted to objections that never matured into returns. Because a return places objections on the journal and forces votes by name, a president who returns a bill accepts a public test that a threat avoids, and a Congress that presses a measure past a stated objection accepts the same test from the other side. The historical totals assembled by the Senate Historical Office for the period from 1789 through the close of the 116th Congress, with 2,584 vetoes in total, 1,518 regular vetoes, 1,066 pocket vetoes, and 112 overrides, describe the visible residue of that bargaining system rather than the measure of the influence the system exerted on bills that changed, narrowed, or disappeared before presentment.

A reader tracking a measure through that system benefits from a record that keeps the procedural evidence separate from the political commentary around the evidence. The study guide to US legislation places the presentment stage inside the longer arc of the series, from introduction and committee gatekeeping through passage in identical form, so that a veto can be read as one event in a sequence of gates rather than as an isolated confrontation. A researcher assembling that sequence for a particular statute can keep your statute notes, citations, and case chronologies together free on VaultBook, holding presentment dates, veto messages, and journal entries in one place as the reconsideration record develops. A student preparing to explain the mechanism under examination conditions can practice and revise US government and civics material on ReportMedic, testing each stage of the sequence against the constitutional text instead of relying on a summary of outcomes.

The diagnostic that the mechanism itself supplies can be applied to any presented measure in the order the Constitution poses the questions. The first question concerns the clock: on what date did presentment occur, and what date completes ten days, Sundays excepted, from that starting point. The second question concerns the return: did the president sign the measure, return the measure with objections inside the period, or allow the period to run without either act. The third question concerns the chamber of origin: if a return occurred, the objections belong on the journal of the chamber where the measure originated, and reconsideration begins in that chamber and in no other. The fourth question concerns the arithmetic in each chamber in sequence: did two thirds of the members present and voting, a quorum being present, agree to pass the bill in the originating chamber, and did two thirds of the members present and voting in the second chamber approve the bill after reconsideration there, with the votes determined by yeas and nays and the names entered on the journal of each chamber. The fifth question concerns the pocket branch, and the question arises only where no return occurred: did adjournment prevent return during the counted period, under arrangements for receipt of messages that Congress had or had not made, so that silence defeated the measure rather than enacting the measure as if signed. A measure that survives each question in order has a determinate legal status, and a measure that fails at any question has a determinate point of failure, and in either case the mechanism, rather than the volume of argument around the measure, has supplied the answer.

In what order does the presentment mechanism test a presented measure?

The test proceeds in a fixed sequence. First comes the clock and the date presentment starts it. Next comes the presidential act, whether signature, return, or silence. Then come the chamber of origin, the vote arithmetic in each chamber in turn, and finally the pocket branch where no return occurred.

Frequently Asked Questions

Q: How many votes are needed for a veto override?

A veto override requires two thirds of the members present and voting in each chamber, a quorum being present, reached in sequence rather than in a joint vote. The chamber where the bill originated reconsiders the returned measure first, and agreement by two thirds of that chamber sends the bill, together with the presidential objections, to the other chamber. Approval by two thirds of the members present and voting in the second chamber enacts the bill into law. The Supreme Court confirmed the present and voting construction in Missouri Pacific Railway Co. v. Kansas in 1919, holding that two thirds of a quorum present suffices. The votes are determined by yeas and nays, and the names of the persons voting for and against the bill are entered on the journal of each chamber respectively, as the constitutional text itself requires.

Q: How often does a veto override succeed?

Overrides have succeeded rarely against the full record of regular vetoes. The veto table of the Senate Historical Office, covering 1789 through the close of the 116th Congress, records 1,518 regular vetoes and 112 overrides, a ratio of about 7.4 percent, or roughly one regular veto in fourteen. The denominator matters to the meaning of the figure. Pocket vetoes cannot be overridden, because no bill and no objections return to Congress for reconsideration, and the Senate Historical Office table counts 1,066 pocket vetoes across the same span. Placing pocket vetoes in the denominator would understate the success rate of the only class of veto that Congress has any opportunity to answer, and careful accounts of override frequency state the source, the span, and the regular veto base alongside the total.

Q: What is a pocket veto compared to a veto override?

A pocket veto and a veto override sit at opposite ends of the return procedure, and the presence or absence of a returned bill separates the two outcomes. A pocket veto occurs when a president neither signs nor returns a bill within ten days, Sundays excepted, and Congress by adjournment prevents return during that period, so the bill does not become law. Nothing returns to either chamber, no objections enter a journal, and no reconsideration can begin. A veto override answers the opposite situation, in which a president returns a bill with objections to the chamber of origin and each chamber then votes, by recorded vote, on whether to enact the bill despite the objections. The pocket veto is absolute in effect because the override procedure has no object on which to operate.

Q: How long does a president have before a veto override question arises?

A president has ten days, Sundays excepted, counted from presentment, within which to sign a bill or return the bill with objections to the chamber where the bill originated. The override question cannot arise before a return occurs, because reconsideration under Article I, Section 7 begins with a returned bill and with objections entered at large on the journal of the originating chamber. Signature inside the period forecloses the question entirely, and inaction inside the period postpones the answer to the expiration of the clock, at which point the bill becomes law as if signed where return stood possible, or fails where adjournment prevented return. The ten day rule therefore fixes both the presidential decision window and the earliest moment at which Congress can face an override vote.

Q: Which chamber votes first on a veto override?

The chamber where the bill originated votes first on a veto override. Article I, Section 7 directs a president who disapproves a bill to return the bill with objections to the chamber in which the bill originated, and that chamber enters the objections at large on the journal and proceeds to reconsider the bill. Only after two thirds of the members present and voting in the originating chamber agree to pass the bill does the measure travel, together with the objections, to the other chamber for reconsideration there. The sequence gives the chamber of origin the first and the framing vote, since a failure to reach two thirds in that chamber ends the override attempt without any vote in the second chamber at all.

Q: Does the president have a line item veto override power?

No federal line item veto exists for a president to exercise, and the override question that the phrasing suggests therefore has no object. Under Article I, Section 7, a president presented with a bill faces the whole measure, and the available responses are signature, return of the entire bill with objections, or inaction under the ten day rule. The Line Item Veto Act of 1996 attempted to add a cancellation power over individual items inside signed laws, and the Supreme Court held the cancellation procedures unconstitutional in Clinton v. City of New York in 1998, on the ground that cancellation in legal effect amended or partly repealed enacted laws outside the procedure the Constitution prescribes. A congressional override answers a veto of a whole bill, and no partial presidential negative exists at the federal level for Congress to answer item by item.

Q: What is a signing statement rather than a veto override?

A signing statement is a presidential statement issued on signing a bill, and the defining contrast with a veto is that the bill becomes law in full. A statement can record constitutional objections to particular provisions or direct how the executive branch will construe provisions of the new law, and the statement accompanies enactment rather than preventing enactment. A signing statement has no formal legal effect on the validity of the statute, and the interpretive weight of a statement has remained contested. A 1993 opinion of the Office of Legal Counsel concluded that signing statements serve legitimate and defensible purposes, including notice to Congress that the executive branch considers a provision unconstitutional in certain applications, while an American Bar Association task force reported in 2006 against the use of signing statements to claim authority to disregard provisions of a signed law.

Q: Do veto threats matter more than a veto override?

Measured by effect on the content of enacted law, veto threats plausibly matter more than overrides, because threats operate on every measure that drafters adjust in anticipation of a return, while an override settles a single returned bill. The mechanism runs through the override threshold itself. A measure that cannot attract two thirds of the members present and voting in both chambers sits exposed to a threatened veto, so sponsors and negotiators revise, narrow, or abandon provisions against the stated objections before presentment. Statements of administration policy state those objections in advance and can record that advisers would recommend a veto, fixing a bargaining position while amendment is possible. The most consequential vetoes in that account are the vetoes never cast, because the bill changed first, and the formal override count captures only the small residue of disputes that bargaining failed to settle.

Q: What record does a returned bill create in the journal of the originating chamber?

A return creates a journal record with three distinct elements. The president returns the bill itself to the chamber where the bill originated, the objections accompany the returned bill as a stated account of the grounds of disapproval, and the chamber enters the objections at large on the journal, reproducing the objections in full rather than in summary. The journal entry preserves the presidential reasoning in the permanent record of the chamber that must act first on reconsideration, so later readers can match the objections against the text that the chamber votes on again. The override votes that follow also enter the journal by name, because Article I, Section 7 requires the votes of both chambers in override cases to be determined by yeas and nays, with the names of the persons voting for and against the bill entered on the journal of each chamber respectively.

Q: How does a bill become law without a signature after a period of presidential inaction?

A bill becomes law without a signature when a president neither signs nor returns the bill within ten days, Sundays excepted, after presentment, and Congress stands in a position in which return would have been possible during that period. The constitutional text assigns inaction a legal consequence in advance, providing that the bill becomes law in like manner as if the president had signed the bill. The result removes any interval of uncertainty in which a measure passed by both chambers and presented in regular course could fail through silence alone, and the rule places the burden of disapproval on an affirmative act of return with objections. The exception defines the boundary of the outcome, because adjournment preventing return during the same period produces a pocket veto instead, under which the bill does not become law.

Q: Why did the Supreme Court invalidate the Line Item Veto Act?

The Supreme Court invalidated the cancellation procedures of the Line Item Veto Act in Clinton v. City of New York, decided on June 25, 1998, by a vote of six to three in an opinion by Justice Stevens. The reasoning turned on the legal character of a cancellation. Once a president signed a bill, the bill was law, and a cancellation of an item inside the law in legal effect amended or partly repealed an enacted statute. The Constitution prescribes a single, finely wrought and exhaustively considered procedure for making law, and amendment and repeal must follow that same procedure, through passage by both chambers and presentment. A statute that authorized a president to cancel items after signature therefore authorized unilateral amendment and repeal, and the Court held that the president may not amend or repeal a statute unilaterally.

Q: What could a president cancel under the Line Item Veto Act before the courts acted?

The Line Item Veto Act of 1996 authorized a president, within five calendar days after signing a law, to cancel in whole any of three categories of provision. The first category was a dollar amount of discretionary budget authority. The second category was an item of new direct spending. The third category was a limited tax benefit. Before cancelling, a president had to determine that the cancellation would reduce the federal deficit, would not impair essential government functions, and would not harm the national interest. A cancellation took effect on receipt of the presidential special message by Congress, and President Clinton sent eleven special messages cancelling eighty two items in total, according to the Congressional Research Service, before the Supreme Court invalidated the procedures in 1998.

Q: How does a veto differ from impoundment of appropriated funds?

A veto and an impoundment operate at different stages of the life of a law and against different objects. A veto acts before enactment, at presentment, when a president returns a passed bill with objections and prevents the text from becoming law unless two thirds of each chamber enacts the bill over the return. Impoundment acts after enactment, when an executive branch decision declines to spend funds that an appropriations law has provided, leaving the statute in force while withholding the expenditure the statute contemplated. Congress answered the impoundment lever in the Congressional Budget and Impoundment Control Act of 1974, which built a separate statutory framework for proposals to rescind or defer appropriated funds. The distinction matters because the override remedy belongs to the veto alone, while impoundment disputes run through the budget statute and through litigation over the duty to execute enacted law.

Q: Why do veto threats change the content of bills before passage?

Veto threats change bills before passage because the override threshold reprices every provision in a measure while amendment is possible. A provision supported by a simple majority in each chamber can pass, travel to presentment, and fail permanently on a return that sponsors cannot answer, since enactment over a veto requires two thirds of the members present and voting in the originating chamber and then in the second chamber. Negotiators who count that arithmetic in advance face a choice between revising the provision toward terms a president can sign and defending the original text through a reconsideration vote that the count says the provision will lose. Administration statements that record an intention to recommend a veto make the arithmetic concrete by naming the provisions at issue, and the resulting revisions enter the text quietly, as drafting decisions, rather than appearing in any record of vetoes cast.

Q: What is a protective return in a dispute over a pocket veto claim?

A protective return pairs a claim that adjournment prevented return with an actual return of the bill and objections, so that the veto stands as a regular veto even if the pocket veto claim fails. President George W. Bush followed that pattern on December 28, 2007, for H.R. 1585, the fiscal 2008 defense authorization bill, issuing a memorandum of disapproval that claimed adjournment had prevented return and invoked the Pocket Veto Case, and simultaneously returning the bill with objections to the Clerk of the House to leave no doubt that the bill was vetoed and to avoid litigation. Congress and the Congressional Research Service treated H.R. 1585 as a regular veto, subject to reconsideration and override. The pattern reflects the unsettled boundary of the pocket branch in an era when both chambers designate agents to receive messages during adjournments and recesses.

Q: Which president issued the most vetoes in the Senate Historical Office table?

Franklin D. Roosevelt issued the most vetoes in the veto table of the Senate Historical Office, which covers 1789 through the close of the 116th Congress. The table credits Roosevelt with 635 vetoes in total, divided between 372 regular vetoes and 263 pocket vetoes, and records nine overrides of Roosevelt vetoes across that span. Grover Cleveland stands second in the same table when both Cleveland presidencies are combined, with 584 vetoes in total, divided between 346 regular vetoes and 238 pocket vetoes, and seven overrides. The Roosevelt total reflects both the length of a presidency that extended across twelve years and the volume of legislation that a Congress active in depression and war presented for signature or return during that tenure.

Q: When did the first successful override of a presidential veto occur?

The first successful override occurred on March 3, 1845, over President John Tyler’s veto of S. 66, an appropriations measure that barred construction of revenue cutters without an appropriation. The House of Representatives voted 126 to 31 to pass the bill over the objections, on the final day of the 28th Congress, according to the House Historian. The date carries a structural lesson alongside the historical one. More than five decades separated the first regular veto, exercised by President George Washington against the Apportionment Act in 1792, from the first completed override, and that interval reflects the height of the two thirds threshold in each chamber as well as the reluctance of early Congresses to press a returned measure to a second vote against stated presidential objections.

Q: How does a statement of administration policy function as an instrument of veto threat?

A statement of administration policy states the position of an administration on pending legislation while the text of a measure is open to change, and the statement functions as a threat instrument when the statement records that advisers would recommend a veto of the measure in a stated form. The formulation matters to the bargaining that follows. A recommendation formula preserves room for negotiation, since drafters can revise the named provisions and argue that the revised text answers the stated objections, while the statement itself has fixed a public bargaining position that the administration cannot abandon without cost. Committees, floor managers, and negotiators between the chambers read such statements as maps of the provisions most exposed at presentment, and the provisions named in a statement attract amendment effort out of proportion to the provisions that pass without comment.

Q: Do absent members change the arithmetic of an override vote?

Absent members change the arithmetic of an override vote because the constitutional threshold runs against the members present and voting, a quorum being present, rather than against the full elected membership of a chamber. The Supreme Court confirmed that construction in Missouri Pacific Railway Co. v. Kansas in 1919, in a unanimous decision by Chief Justice White concerning enactment of the Webb-Kenyon Act over a veto, holding that two thirds of a quorum present suffices. Absences therefore lower the number of affirmative votes that two thirds requires, while the quorum requirement sets a floor below which a chamber cannot conduct the vote at all. A report that converts an override margin into a fraction of all members, without accounting for attendance on the day of the vote, misstates the constitutional test that decided the outcome.

Q: What would a future federal line item veto require after the invalidation of the statute?

A future federal line item veto would require a constitutional amendment, because Clinton v. City of New York closed the statutory route in 1998. Congress attempted to create the power by ordinary legislation in the Line Item Veto Act of 1996, and the Supreme Court held that the cancellation procedures the Act created violated the Presentment Clause, since a cancellation in legal effect amended or partly repealed a law that signature had already enacted. A statute cannot confer a power that the Court has located outside the procedure the Constitution prescribes, so a later Congress repeating the same device in new statutory language would meet the same holding. An amendment could write an item veto into the constitutional text itself, as state constitutions have done for most governors, and amendment is the only route that changes the rule the 1998 decision applied.