On December 12, 1985, President Ronald Reagan signed the Balanced Budget and Emergency Deficit Control Act of 1985, Public Law 99-177, the statute known by the names of its Senate sponsors as Gramm-Rudman-Hollings. Enactment closed a year in which the fiscal 1985 unified deficit in the Office of Management and Budget Historical Tables had reached $212 billion, equal to 5.1 percent of gross domestic product, and in which a budget resolution had assumed a lower figure while requiring $55.5 billion of reduction legislation that Congress never enacted. A legislature tried to legislate an outcome. Congress wrote a descending schedule of maximum deficit amounts into law, attached an automatic cancellation of budgetary resources to any estimated breach of the schedule, and presented the arrangement as a path to a balanced budget on a date fixed in advance.

The sponsors who carried the measure gave that ambition a cross party form. Senator Phil Gramm, a Republican of Texas, Senator Warren B. Rudman, a Republican of New Hampshire, and Senator Ernest F. Hollings, a Democrat of South Carolina, lent their names to a discipline that restrained congressional discretion and executive discretion alike. The Senate added their amendment to H.J.Res. 372, a measure raising the statutory debt limit to $2.079 trillion, so the deficit regime traveled on borrowing authority that the government required rather than on a freestanding bill that either chamber could postpone. Conference and an amendment exchange between the chambers produced the final text, the House approved the measure by 271 to 154, the Senate approved the measure by 61 to 31, and the signature followed one day after congressional agreement on December 11, 1985.

The schedule at the center of the statute descended in stated steps. The maximum amount was $171.9 billion for fiscal 1986 and $144.0 billion for fiscal 1987, followed by $108.0 billion, $72.0 billion, and $36.0 billion, with balance at zero for fiscal 1991. A margin of $10 billion applied in the intervening years, so an estimate could exceed a printed ceiling by that amount without triggering cancellation. The trigger itself ran on an estimate prepared before the fiscal year began, assembled through reports from the Office of Management and Budget, the Congressional Budget Office, and the Comptroller General, and converted into a presidential sequestration order. Sequestration meant automatic cancellation of budgetary resources, split equally between defense and nondefense categories and applied as a uniform percentage within each category, while more than half of federal spending stood exempt or protected outside the base.

One of the authors of the act, quoted in the Congressional Research Service report on statutory budget controls, explained the design as a threat rather than as a preferred method of budgeting, observing that the objective was never to trigger the sequester and that the objective was to have the threat of the sequester force compromise and action. Fixed numbers prevented the destination from drifting during annual bargaining, the uniform percentage prevented officials from ranking programs at the moment of failure, and the exemptions recorded the boundaries of the coalition that enactment required. The same features built weaknesses into the law, because a forecast could move the trigger, a protected base could concentrate the percentage on unprotected programs, and a legislature that had written a ceiling by statute retained the power to rewrite the ceiling by statute.

The constitutional weakness appeared on the day of signing. The presidential signing statement recorded the view that the act was constitutionally defective because of the supervisory role assigned to the Comptroller General, an officer removable by Congress who evaluated the estimates on which the presidential order rested. Codification placed the act principally at title 2 of the United States Code, chapter 20, section 901 and following, and the statute amended the budget framework enacted in 1974, including elimination of the requirement of a second concurrent budget resolution. The law therefore entered force as a complete system with a contested engine and a descending table with a balance date.

Origins: deficit politics before the statute

The Balanced Budget and Emergency Deficit Control Act of 1985 began in a fiscal position that neither chamber could describe as temporary drift. The fiscal 1985 unified deficit in the Office of Management and Budget Historical Tables reached $212 billion, equal to 5.1 percent of gross domestic product. A shortfall of that scale formed the political setting for every procedural choice that followed. The figure mattered not only as a measure of borrowing but also as evidence that the annual budget process, operating through its ordinary sequence of resolutions, authorizations, and appropriations, had failed to produce a path toward balance that commanded agreement in either chamber or at the White House.

The fiscal 1986 budget resolution sharpened that failure into a specific, documented gap between promise and performance. The resolution assumed a deficit of $171.9 billion for fiscal 1986 and required $55.5 billion of deficit reduction legislation to make the assumed figure real. That legislation was never enacted. The distance between the two numbers therefore described more than an estimating problem. Congress had adopted a fiscal plan whose central figure depended on laws that Congress did not pass, and the plan contained no mechanism capable of closing the distance on its own. When proponents of a new statute argued that ordinary process had visibly failed, the unenacted $55.5 billion supplied the clearest exhibit for the argument.

That failure carried particular weight because the budget process itself was a comparatively deliberate construction. Congressional Budget Act of 1974 had created the Congressional Budget Office and had built a budget resolution process through which Congress was to state its fiscal intentions in an orderly form before legislating within them. The 1985 act amended that 1974 framework rather than replacing it, and the amendment overlaid the older process with fixed deficit targets and an enforcement device, sequestration, that the older process lacked. The origins of the 1985 statute thus lay inside an existing institutional design whose weakness had been identified with precision: a resolution could announce a deficit figure, and could even specify the amount of reduction legislation needed to reach it, without compelling any legislator to vote for the taxes, spending restraint, or combination of the two that the figure required.

The absence of compulsion shaped the remedy that the sponsors selected. Rather than write another resolution restating an intention, the sponsors designed a law under which failure to reach a stated deficit maximum would produce automatic cancellation of budgetary resources. The enforcement device, described in the statute as sequestration, was meant to operate without a further affirmative vote at the moment of failure. That design choice reversed the burden of action that had defeated the fiscal 1986 resolution. Under the older sequence, inaction preserved spending and revenue law unchanged and allowed the deficit to settle wherever economic conditions and prior law placed it. Under the proposed sequence, inaction carried a consequence specified in advance, and legislators who wished to avoid that consequence had to assemble a majority for an alternative.

The three sponsors gave the proposal a political character that a single party sponsorship could not have supplied. Senator Phil Gramm was a Republican of Texas. Senator Warren B. Rudman was a Republican of New Hampshire. Senator Ernest F. Hollings was a Democrat of South Carolina. The sponsorship crossed party lines, with two Republicans and one Democrat lending their names to a measure that restrained a Republican administration’s fiscal room as firmly as it restrained congressional discretion. That composition mattered in a debate where deficit reduction could easily have been framed as an attack by one party on the priorities of the other. A Democrat’s name on the amendment signaled that concern over the deficit had a constituency beyond the majority party of the Senate, and the two Republican names signaled that the concern extended inside the party of the President who would have to sign or veto the result.

Cross party sponsorship also reflected the distribution of the burden that the statute contemplated. The mechanism divided a required reduction equally between defense and nondefense categories, with a uniform percentage reduction applied within each category, so the pain of failure was not assigned to one side of the budget alone. A proposal carrying only Republican sponsorship might have been read as an instrument against domestic programs; a proposal carrying only Democratic sponsorship might have been read as an instrument against defense. The names Gramm, Rudman, and Hollings allowed the measure to be presented instead as a discipline imposed on the budget as a whole, and the equal division between categories reinforced that presentation in the substance of the design.

One of the act’s authors, quoted in the Congressional Research Service report on statutory budget controls, stated the purpose behind the threatening design in terms that ruled out any reading of the sequester as a preferred outcome. The author said: “It was never the objective of Gramm-Rudman to trigger the sequester; the objective of Gramm-Rudman was to have the threat of the sequester force compromise and action.” The sentence explained the architecture more directly than any summary of the targets could. The automatic cancellation existed to make negotiated agreement the more attractive course, and its severity was a feature of the bargaining structure rather than an expression of confidence that across the board reductions represented sound policy in themselves.

That stated intent also clarified why the sponsors accepted a mechanism that treated dissimilar programs through a common percentage. A sequester computed as a uniform percentage reduction within the defense category and within the nondefense category made no claim to rank programs by merit, efficiency, or need. Its function lay elsewhere. Because the reductions would fall mechanically, without the selective judgments that ordinary appropriations and authorizing legislation supply, both parties and both branches had reason to prefer a negotiated package in which tradeoffs could be chosen deliberately. The threat therefore disciplined the negotiation precisely because the fallback was indiscriminate, and the sponsors’ own description of the objective treated the fallback as a penalty for failure rather than as the policy the statute sought to deliver.

The fiscal arithmetic of 1985 left the sponsors little room to present their proposal as a modest adjustment. A unified deficit of $212 billion at 5.1 percent of gross domestic product stood far from the balanced outcome that the statute’s title announced, and the path from that starting point to balance required annual reductions large enough to force recurring confrontation with the same choices that the fiscal 1986 resolution had avoided. The target sequence that the enacted law adopted moved the maximum deficit downward year by year until it reached zero, which meant that the political difficulty was not confined to a single vote in 1985. Each succeeding year reopened the question under a lower ceiling, and each succeeding Congress and administration inherited a test that grew harder as the permitted figure fell.

The vehicle chosen for the amendment determined whether that agreement would ever be tested on the floor. A freestanding deficit bill could have been delayed, amended beyond recognition, or defeated without collateral consequence, and the record of the unenacted fiscal 1986 reduction legislation demonstrated how readily fiscal intentions dissolved when a vote could be postponed. The sponsors instead attached their proposal to legislation that Congress had to complete. The Senate added the Gramm-Rudman-Hollings amendment to H.J.Res. 372, a measure raising the statutory debt limit to $2.079 trillion, the borrowing ceiling governed by debt ceiling law. The pairing joined a new enforcement regime to an increase in borrowing authority that the federal government required, and the joining was deliberate rather than incidental.

Why was the measure attached to a debt limit bill?

The Senate added the Gramm-Rudman-Hollings amendment to H.J.Res. 372 because the debt limit increase to $2.079 trillion had to pass, giving the deficit measure a vehicle that could not easily be set aside and forcing both chambers to confront targets they had failed to reach through the ordinary budget process.

The procedural logic of the vehicle extended beyond the Senate. Because the amendment traveled on H.J.Res. 372, the House confronted the deficit proposal as part of a measure whose failure would leave the debt limit unchanged, rather than as a separate question that could be deferred to a more convenient session. Conference negotiations between the chambers then proceeded under the pressure that the underlying measure supplied. The brief for the series frames the episode as the first prominent modern instance of the debt limit carrying unrelated fiscal legislation, and the pattern recurred in later decades, with the 1987 reaffirmation of the same statute again enacted on a debt limit increase. The 1985 pairing therefore established a precedent of practical significance: a must pass borrowing measure could serve as the legislative carrier for a budget process reform that might not have survived as a freestanding bill.

The political logic matched the procedural logic. Voting against the debt limit increase because of the attached amendment required a member to accept responsibility for blocking the borrowing authority, while voting for the increase required acceptance of the targets and the sequester as the price of passage. That structure compressed the choices available to opponents, who could no longer defeat the deficit measure in isolation and then support the debt limit on its own terms. The amendment’s sponsors gained leverage from the compression, and the leverage was greatest precisely because the deficit proposal was controversial enough to face an uncertain fate on a direct vote. The vehicle did not manufacture agreement on fiscal policy, but it raised the cost of expressing disagreement to a level that many members were unwilling to pay.

The route to enactment reflected the difficulty that the vehicle was chosen to overcome. The bill was agreed on December 11, 1985, after conference and an amendment exchange between the chambers, a sequence in which the House and Senate traded amended versions back and forth rather than settling the text through a single clean adoption. That exchange, sometimes described through the shorthand of ping pong between the chambers, testified to the number of contested points that survived the committee and floor stages: the level and timing of the targets, the division of reductions between defense and nondefense programs, the exemptions that would shield particular programs from sequestration, and the assignment of the reporting and certification duties that made the mechanism automatic. Each of those questions touched a constituency prepared to defend the existing distribution of spending, and the amendment exchange gave both chambers repeated opportunities to adjust the balance among them.

Final passage, when the votes came, produced majorities that were comfortable without approaching unanimity. The House approved the measure by 271 to 154. The Senate approved it by 61 to 31. Those margins recorded a coalition broad enough to enact a statute over determined opposition, and they also recorded the scale of the dissent that the enforcement concept attracted even on a must pass vehicle. More than a third of the House and a third of the Senate voted against a bill that combined a necessary debt limit increase with the deficit regime, which indicated that the sequester idea itself, and not merely the inconvenience of the vehicle, drove a substantial share of the opposition. The affirmative majorities nonetheless held in both chambers on the same text, and the conference agreement survived the amendment exchange without losing the automatic mechanism at its center.

President Ronald Reagan signed the measure on December 12, 1985, one day after congressional agreement. The signature converted the sponsors’ amendment into the Balanced Budget and Emergency Deficit Control Act of 1985, Public Law 99-177, published at 99 Stat. 1037. The statute took its popular name, Gramm-Rudman-Hollings, from its three Senate sponsors, a naming practice that preserved the cross party character of the sponsorship in the label by which the law entered public and scholarly discussion. Codification placed the act principally at title 2 of the United States Code, chapter 20, section 901 and following, within the body of law governing Congress and the budget process that the statute amended.

The placement of the codification carried meaning of its own. By amending the Congressional Budget Act of 1974, the new statute inserted its targets and its enforcement procedure into the same legal framework that housed the budget resolution process, so that the two systems operated together rather than in separate statutory worlds. Among the amendments to the 1974 act, the 1985 statute eliminated the requirement of a second concurrent budget resolution, streamlining the older calendar at the same time that it tightened the consequences attached to the figures the remaining resolution process produced. The reform thus subtracted a procedural step while adding a substantive obligation, a combination that captured the sponsors’ diagnosis: the difficulty with congressional budgeting lay not in a shortage of procedural stages but in the absence of a penalty when the stages failed to yield the announced result.

The signing did not reflect unqualified presidential confidence in the statute’s design. President Reagan’s signing statement itself stated the view that the act was constitutionally defective because of the supervisory role assigned to the Comptroller General, a position recorded at 21 Weekly Comp. Pres. Doc. 1491 (1985). The statement flagged, at the moment of enactment, the precise vulnerability that litigation would test within the following year: the Comptroller General, as head of the General Accounting Office, evaluated the deficit and sequestration reports of the Office of Management and Budget and the Congressional Budget Office and reported the conclusions on which the President’s sequestration order rested, yet the Comptroller General was an officer removable by Congress. A President signing a law while announcing doubt about the constitutionality of its central mechanism presented an unusual spectacle, and the statement ensured that the doubt formed part of the statute’s record from its first day in force.

That early warning deserves emphasis in any account of the statute’s origins because the doubt was structural rather than rhetorical. The automatic character of the mechanism depended on placing the decisive estimate and certification outside the discretion of officials who might soften the result, and the sponsors had assigned that function to the Comptroller General as an officer positioned between the branches. The constitutional objection held that functions amounting to execution of the laws could not be vested in an officer whom Congress could remove. The signing statement placed the executive branch on record against the arrangement at the same time that the executive branch accepted the statute as a whole, which meant that the law began its life with its enforcement design already contested by one of the branches charged with operating it.

The origins of the act therefore contained, in compressed form, the tensions that governed its entire career. A deficit of $212 billion in fiscal 1985, at 5.1 percent of gross domestic product, supplied the pressure. A budget resolution that assumed $171.9 billion for fiscal 1986 and required $55.5 billion of reduction legislation that never arrived supplied the proof that exhortation without enforcement had failed. Three sponsors drawn from both parties supplied a coalition that could present discipline as a shared obligation rather than a partisan instrument. A debt limit increase to $2.079 trillion supplied the vehicle that made postponement costly. Conference and an amendment exchange supplied the final text, majorities of 271 to 154 in the House and 61 to 31 in the Senate supplied the votes, and a presidential signature on December 12, 1985 supplied the enactment, accompanied by a signing statement that questioned the constitutionality of the mechanism the signature brought into force.

The author’s statement of intent, preserved through the Congressional Research Service, supplied the interpretive key that the sponsors themselves offered for the whole construction. The sequester was a threat held in reserve, designed to force compromise and action by making the alternative to compromise sufficiently unattractive to every participant in the budget process. Read against that statement, the statute’s origins appear as an exercise in institutional leverage rather than in fiscal arithmetic alone. The sponsors could not command agreement on which programs to reduce or which revenues to raise, and the fiscal 1986 record showed that announcing the necessary total accomplished nothing by itself. The sponsors therefore legislated a consequence, attached the consequence to a measure that had to pass, and relied on the prospect of automatic, indiscriminate cancellation to accomplish what resolution language had not. Whether the device would work as designed, and whether its design could survive constitutional scrutiny, were questions that enactment posed rather than answered.

The design: fixed targets and an automatic trigger

The design of the act rested on linked decisions about numbers, timing, and compulsion. Congress set a declining path for the federal deficit in statute, arranged for an official estimate of the coming fiscal year to be tested against that path before the year began, and prescribed an automatic cancellation of budgetary resources if the estimate breached the permitted level by more than a stated margin. Each element altered the incentives facing the President and Congress. A fixed path replaced negotiation over the destination with negotiation over the means of reaching a destination already written into law. An estimate as the trigger moved the decisive moment forward, before spending for the year had fully unfolded and before an audited result could be known. A uniform percentage as the method of reduction removed official discretion over which accounts would bear the burden inside each broad category. The mechanism therefore constructed a calendar, a test, and a consequence, and tied the three together so that inaction itself had a programmed result.

The statutory path began with a maximum deficit amount of $171.9 billion for fiscal 1986. For fiscal 1987 the maximum amount fell to $144.0 billion. For fiscal 1988 the ceiling was $108.0 billion, and for fiscal 1989 it was $72.0 billion. For fiscal 1990 the amount declined to $36.0 billion, and for fiscal 1991 the target was zero, a balanced position expressed as a maximum deficit amount of no dollars. The path thus asked the federal government to move from a deficit near the level contemplated for fiscal 1986 to complete balance in five further steps. The numbers were absolute rather than relative. The statute did not define success as a share of national output, as an improvement over a prior year, or as a reduction adjusted for economic conditions. It named a dollar ceiling for each fiscal year and attached consequences to an estimate that rose too far above that ceiling. That choice gave the schedule clarity, because a dollar figure either had or had not been crossed, and it exposed the schedule to every development that moved federal receipts and outlays in dollar terms, from legislation to economic performance different from the forecast to the timing of transactions across fiscal years.

The spacing of the steps carried its own meaning. The first reduction in the maximum amount, from $171.9 billion for fiscal 1986 to $144.0 billion for fiscal 1987, was $27.9 billion. Each later step through fiscal 1991 was $36 billion, from $144.0 billion to $108.0 billion, from $108.0 billion to $72.0 billion, from $72.0 billion to $36.0 billion, and from $36.0 billion to zero. The regularity of the $36 billion increments presented the descent as orderly and predictable, a staircase with equal treads after the opening move. Predictability served a bargaining purpose, because departments, committees, and the executive branch could measure proposals against a known annual requirement rather than against a shifting political judgment. The equal steps also concealed an asymmetry in difficulty, because a $36 billion reduction from a lower base was not the same legislative task as the same interval from a higher base, and the final interval demanded elimination of the entire remaining amount.

Beside most of the ceilings stood a margin of $10 billion. The margin was the amount by which the estimated deficit could exceed the target without triggering a sequester. Fiscal 1986 had no margin, and fiscal 1991, the balance year, also had no margin. Fiscal 1987, fiscal 1988, fiscal 1989, and fiscal 1990 each carried the $10 billion allowance. In practical terms, the operative trigger levels in those intervening years were $154 billion for fiscal 1987, $118 billion for fiscal 1988, $82 billion for fiscal 1989, and $46 billion for fiscal 1990. An estimate at the ceiling required no cancellation, and an estimate above the ceiling but within the margin also required no cancellation. Only an estimate beyond the margin activated the prescribed consequence. The margin acknowledged that the test rested on a projection and that a projection could miss by a modest amount without proving that the underlying budget plan had failed. It built a tolerance band into an otherwise rigid rule and reduced the chance that a small forecasting difference would set off a large administrative operation. At the same time, the margin loosened the announced ceiling in every year in which it applied, because the legal test permitted the target plus $10 billion before compulsion began, and bargainers aiming to avoid an automatic order had reason to aim for the trigger level rather than for the lower figure that attracted public attention.

The most consequential choice in the mechanism was the identity of the event that triggered action. The trigger was an estimate produced before the fiscal year began, described by the August snapshot, and not the audited outcome for that year. The law did not wait until receipts had been collected, outlays had been recorded, and the final deficit had been measured. It asked officials to project the coming year, compare the projection with the statutory maximum and margin, and order cancellations on the basis of that comparison. Speed was the principal gain. A remedy founded on a completed outcome would have arrived after the fiscal year in question had closed, when the budgetary resources for that year could no longer be rearranged in the same manner. The estimate allowed the consequence to fall inside the year being controlled and allowed the threat of that consequence to influence decisions while those decisions remained open.

The price of speed was dependence on a forecast. An estimated deficit embodied assumptions about economic growth, receipts, outlays, and the effects of law for a year that had not begun. Different assumptions could move the estimate across the trigger line without any change in enacted policy, and changes in enacted policy could be scored in ways that moved the estimate without an equivalent movement in the eventual outcome. The statute nevertheless treated the August figure as the operative fact for purposes of the order. That treatment gave forecasting institutions a central role in fiscal discipline and gave every participant an incentive to contest, select, or shape the assumptions behind the number. A target set on a forecast disciplined plans rather than results. It rewarded a plan that appeared compliant at the snapshot moment and penalized a plan that appeared noncompliant at that moment, regardless of how later events rearranged receipts and spending.

A snapshot also created a boundary, and boundaries invite attention to transactions that fall on one side or the other. Because the test was applied at a fixed point and to a defined fiscal year, a transaction counted differently depending on when it was recorded and how it was scored for that year. The record associated with the statute included the use of one-time asset sales to generate paper savings against targets, including sales of government loan assets, and the shifting of payment dates across the fiscal year boundary so that an outlay fell in one year rather than another. Such devices acquired value from the structure of the test even when they did little to alter the longer fiscal position, because the legal consequence depended on the estimated position for a single year as captured at a particular moment. A design that measured a completed multi-year outcome would have reduced the payoff to moving a receipt or a payment by a short interval, while a forecast for a single year at an August snapshot increased that payoff.

The reporting calendar translated that design into a sequence of institutional acts spread across late summer and autumn. In August the Office of Management and Budget and the Congressional Budget Office issued a joint deficit and sequestration report. The joint character of the opening report produced a common starting point for the calculations that followed. On August 25 the Comptroller General, the head of the General Accounting Office, issued a report evaluating the report from the Office of Management and Budget and the Congressional Budget Office. On September 1 the President issued an initial sequestration order. That initial order took effect in connection with the final order on October 15, after updated information had passed through a second round of reports. On October 5 the Office of Management and Budget and the Congressional Budget Office issued updated reports. On October 10 the General Accounting Office issued its updated report. On November 15 the General Accounting Office issued a compliance report. The calendar therefore separated estimation, evaluation, initial presidential action, revision, final effect, and compliance review into distinct dated stages, each with a defined institutional author.

The sequence also distributed institutional responsibility. The August 25 evaluation placed a reviewer outside the two offices that had prepared the joint report before presidential action, the interval between September 1 and October 15 left room for the updated reports to revise the basis for action, and the November 15 compliance report assigned the General Accounting Office a review of whether the action taken conformed to the requirements after the final order. The process was automatic in its consequence but elaborate in its choreography, and the elaboration mattered because the legitimacy of an indiscriminate reduction depended in part on a visible chain of reports, evaluations, and orders.

What happened when the estimated deficit exceeded the target by more than the margin?

The President issued a sequestration order that canceled budgetary resources by a uniform percentage within the defense category and within the nondefense category, with the total reduction split equally between the two categories, in the amount the reports calculated was needed to reach the target.

Sequestration, as a mechanical term, meant the automatic cancellation of budgetary resources. The object canceled was not an abstraction called spending in general. The cancellation reached budget authority and related resources, the legal capacity provided through appropriations and other budgetary arrangements for obligations and outlays. Cancellation removed resources that otherwise would have been available under the governing budgetary arrangements. The statute prescribed the arithmetic by which the required total reduction was distributed. The total was split equally between two categories, defense and nondefense, so that one half of the required reduction fell on the defense category and one half fell on the nondefense category. Within each category, the reduction was computed as a uniform percentage. A program or account inside a category therefore faced the percentage applicable to that category rather than a percentage selected account by account according to an official ranking of merit, urgency, or efficiency.

The equal split between defense and nondefense expressed a political settlement inside the arithmetic. Neither broad category was assigned a larger share of the automatic burden, and neither could present the automatic consequence as a device aimed principally at the other. That symmetry made the threat mutually uncomfortable and supported the argument that the order was a penalty for collective failure rather than a disguised appropriations decision favoring one set of priorities. Symmetry also limited the ability of the mechanism to reflect the composition of the deficit or the location of the policies that had produced the excess estimate, because the required reduction followed the statutory division rather than an inquiry into which category had grown or contained resources that could be withdrawn with the least disruption.

The uniform percentage inside each category carried the same logic further. Uniformity removed the need for the President, the estimating offices, or the Comptroller General to rank programs at the moment of failure, and it removed a large field of discretion in which bargaining could have reopened the very choices the automatic process was meant to foreclose. Every covered resource in a category received the same proportional treatment, so the method could be described as impartial in form. The impartiality was formal rather than substantive. Programs differed in fixed obligations, in the timing of their outlays, and in their ability to defer purchases or activities. A single percentage translated those differences into unequal practical effects while preserving equal arithmetic treatment. By refusing priorities, the design avoided an official statement that some activities mattered more than others, but the refusal did not abolish priority in practice. It transferred the burden of adjustment to managers and beneficiaries operating under unlike constraints, after the statute had declined to distinguish among them.

The refusal to rank also protected the credibility of the threat. A selective automatic cut would have required a standing list of preferred and unpreferred activities, and such a list would have become an object of legislation and lobbying long before any trigger date arrived. A uniform percentage made the consequence harder to tailor in advance and therefore harder for individual constituencies to discount as a penalty that would fall mainly elsewhere. The breadth of exposure was part of the deterrent, even though the resulting cancellations embodied no judgment that the resources removed were the least valuable in the category.

The statute provided for suspension of the mechanism in time of war or recession, acting through a joint resolution. The recession branch rested on growth tests stated in the statute. One test looked to projected real growth and treated two consecutive quarters of negative projected real growth as the relevant condition. The other test looked to actual real growth and treated actual real growth below 1 percent for two consecutive quarters as the relevant condition. The projected test allowed the suspension inquiry to begin on an expected contraction before a complete record of actual performance existed, while the actual test supplied a separate route founded on measured growth. War and recession were treated as exceptional states in which the normal discipline of the target path could be set aside through the prescribed legislative form, and the joint resolution requirement kept suspension from resting on a unilateral declaration by the President or on an administrative finding by an estimating office.

The limits on suspension were as instructive as the permission. A suspension could not undo a sequestration order already in effect. Relief operated forward from the suspension rather than backward across an order that had already taken legal effect. That restriction preserved the finality of completed action and raised the cost of delay in seeking relief, because actors who expected war or recession conditions to justify suspension had to secure the joint resolution before an order crossed into effect if they wished to prevent the cancellations entirely. Once the order was in effect, the resources already canceled under that order remained subject to the action taken. The escape route was therefore real but narrow in time.

For fiscal 1986, the first year under the statute, a special rule replaced the ordinary scale and timing of the consequence. The sequester was capped at $11.7 billion. The order took effect on March 1, 1986, rather than on the October schedule associated with the regular sequence for a coming year. The March date fell partway through a fiscal year that had begun the previous October, and the statute applied a seven-twelfths proration to the year. Proration recognized that only seven of the twelve months remained when the order took effect and scaled the operation to the remaining portion instead of pretending that a full year of budgetary execution lay ahead. The cap limited the total automatic reduction for the transitional year to $11.7 billion, far below the distance between the $171.9 billion maximum amount for fiscal 1986 and the larger deficit that an uncapped application of the ordinary logic might have pursued. The special rule thus acknowledged that the statute had been enacted after the fiscal year had begun, and that forcing the full annual path in the remaining months would have compressed the reduction into a shorter period than the regular design contemplated.

The realized reductions under the fiscal 1986 rule were approximately 4.3 percent for nondefense programs. That percentage gives a concrete sense of the difference between the headline architecture and the first application. The architecture promised a fixed descent to balance enforced by automatic cancellation. The first application operated under a cap, a delayed effective date, and a proration formula, and it produced a mid-single-digit percentage reduction on the nondefense side rather than an immediate correction of the whole excess over the target. The moderation was built into the statute for that year and did not depend on a later decision to soften the order. Congress had written the transition terms in advance, limiting the inaugural sequester while establishing the precedent that an order could in fact issue and cancel resources.

Viewed as a whole, the design sought credibility through rigidity and deterrence through discomfort. Fixed dollar targets prevented the destination from drifting during annual bargaining, the margin supplied a limited tolerance for projection error, and the August estimate made the test occur early enough to affect the year under examination. The reporting calendar gave the test institutional form through the joint August report, the August 25 evaluation, the September 1 initial order, the October 15 final effect, the updated reports on October 5 and October 10, and the November 15 compliance report. Sequestration then imposed an equal division between defense and nondefense and a uniform percentage within each category, a formula that no participant could easily mistake for a considered ranking of public purposes. Suspension for war or recession supplied an exception for extraordinary conditions, bounded by growth tests, by the joint resolution form, and by a bar on undoing an order already in effect, and the fiscal 1986 cap and proration supplied a separate, transitional moderation. The mechanism was strongest as a threat held in reserve and most exposed when its formulas had to govern real resources, because the qualities that made the threat clear, equal, and automatic were the same qualities that prevented it from discriminating among forecasts, programs, and circumstances.

Exemptions and the protected half of the budget

The Balanced Budget and Emergency Deficit Control Act of 1985 did not place the entire federal budget within reach of its automatic cancellation mechanism. Before any percentage could be calculated, the statute drew a boundary around a large group of programs and placed that group outside the process. Sequestration meant the automatic cancellation of budgetary resources, computed as a uniform percentage reduction within each of two categories, defense and nondefense, with the total required reduction split equally between the two categories. The exemptions determined the base to which that percentage applied. A program outside the base contributed nothing to a sequester, however large the required reduction became. A program inside the base absorbed the full calculated percentage, and the percentage itself grew larger because so much spending had been removed from the calculation at the outset. The protected share of the budget therefore shaped the arithmetic of every order and set practical limits on what the process could accomplish when an estimate exceeded a target by more than the permitted margin.

The Congressional Research Service account of the statute identified a group of programs as fully exempt from sequestration. The exempt programs were Social Security; Medicaid; veterans’ compensation; veterans’ pensions; Aid to Families with Dependent Children; Supplemental Security Income; the Special Supplemental Food Program for Women, Infants, and Children, known as WIC; food stamps; the postal service fund; and the earned income tax credit. The exemption was categorical rather than conditional. A fully exempt program did not face a reduced percentage, a delayed percentage, or a percentage applied only after other accounts had been exhausted. Exemption removed the program from the sequestration base altogether, and an order issued under the act left its budgetary resources untouched by that process.

Net interest received protection in operation as well. The Congressional Research Service description treated net interest as exempt in operation, a treatment that placed the cost of servicing the federal debt alongside the named exempt programs in practical effect. Interest obligations arose from borrowing that had already occurred, and the cancellation mechanism operated on budgetary resources for programs and activities rather than rewriting obligations created by past debt. The working design of the statute therefore accepted that interest payments would continue outside the sequester calculation. The result was a base for reductions that excluded both the major benefit programs named in the exemption list and the interest cost generated by earlier deficits. More than half of federal spending was exempt or protected, which concentrated the cuts on the remaining base. That proportion governed every later dispute about percentages, because a percentage figure in a sequestration report described a reduction from the unprotected remainder rather than a reduction from federal spending as a whole.

Social Security occupied a distinctive position within the protected group because the 1985 act gave the program two forms of separation from the enforcement process. Social Security appeared among the fully exempt programs, and the same act placed Social Security off budget beginning in 1986, a feature of the record summarized in Senate Report 105-3. Exemption protected the program from cancellation under a sequestration order. Off budget placement reinforced a separate budgetary status for the program from 1986 onward. Together the provisions marked Social Security as the program least connected to the enforcement machinery of the act. The choice also had arithmetic significance. Removal of a small account from a base alters a denominator at the margin. Removal of a program of that scale leaves the remaining accounts facing percentages calculated without the breadth that a larger base would have supplied.

Below the fully exempt tier, the act created a second tier of protection for programs that remained within reach of sequestration under a strict ceiling. The special rule programs were Medicare; veterans’ medical care; community health centers; migrant health centers; and Indian health facilities and services. Reductions for those programs were capped at 1 percent in 1986 and at 2 percent in subsequent years. The cap transformed the nature of exposure. An ordinary unprotected program faced whatever uniform percentage the calculation produced for its category. A special rule program faced a known maximum set in the statute, bearing no direct relationship to the size of the overall shortfall. If the uniform percentage for a category reached 8 percent or 10 percent or more, the capped programs contributed at the 1 percent or 2 percent level fixed for the year in question, and the difference had to be found elsewhere within the base or accepted as a limit on the total resources an order could reach. Health related programs dominated the tier, with Medicare as the largest member, and the tier system meant that the act graded protection rather than adopting a single rule. Full exemption served one set of programs, a percentage ceiling served another set, and unprotected status governed the remainder.

Two further programs received specialized treatment outside both the exemption list and the cap structure. Guaranteed student loans and child support enforcement had special calculation rules under the act, as the Congressional Research Service account recorded. A special calculation rule acknowledged that a straightforward uniform percentage cancellation did not fit every account in the same way. Some budgetary resources represent direct provision of a service by an agency, while other resources support guarantees, enforcement activity, or transactions whose budgetary effect follows a structure that a single percentage cannot capture cleanly. The act handled that mismatch by specifying a separate method of calculation for those programs rather than forcing the standard formula onto accounts whose operation the formula described poorly. Each departure from the uniform formula narrowed or complicated the base a little further.

The arithmetic consequence followed from elementary division applied to an unusually consequential denominator. A sequester began with a required dollar amount, derived from the distance between an estimated deficit and the statutory target after allowance for the margin. That amount was split equally between defense and nondefense. Within each category, the amount assigned to the category was divided across the budgetary resources that remained subject to cancellation, after exempt programs had been removed and after capped programs had been limited to the contribution their ceilings permitted. The quotient became the uniform percentage for the category. Every dollar of exempt spending reduced the denominator without reducing the numerator. Every capped program that contributed less than the uniform percentage would have required left a further share of the numerator to be collected from uncapped accounts. The percentage for unprotected programs therefore rose from two directions at once, once because the total base had shrunk and again because protected programs inside the categories contributed below the calculated figure. Across the budget as a whole, the burden was anything but uniform.

The operating record illustrated how quickly that percentage could climb. For fiscal 1986, a special rule capped the sequester at $11.7 billion, and the order took effect on March 1, 1986 under a proration covering seven twelfths of the year. Even under that capped and partial year arrangement, the realized reduction percentages reached approximately 4.3 percent for nondefense programs and 4.9 percent for defense programs. Those figures represented reductions from the unprotected base only, since exempt programs contributed nothing and capped programs contributed at the 1 percent level set for 1986. For fiscal 1988, the Congressional Budget Office calculated uniform reduction percentages of 10.4 percent for defense and 8.7 percent for nondefense against a threatened sequester framed around $23 billion in outlay reductions split evenly between the categories. By the fiscal 1991 calculation, the Office of Management and Budget final sequestration report in October 1990 estimated reductions that would have required defense outlay reductions of 34.7 percent and nondefense reductions of 31.6 percent. As the required amount grew, the unprotected remainder had nowhere to spread the burden, and the percentage rose toward levels that described wholesale cancellation of activity rather than trimming at the edges.

Defense bore a particular weight within that arithmetic because the statute assigned half of every required reduction to the defense category regardless of the distribution of exempt spending across the budget. The equal split was a fixed feature of the original design rather than a proportion recalculated to reflect the size of each category after exemptions. The nondefense category contained the large exempt benefit programs, which removed much of its spending from exposure while leaving its assigned share of the reduction to be collected from a reduced pool of exposed resources. Defense had to produce its full half of the required amount from the defense resources that remained subject to cancellation. The 1987 repair acknowledged the pressure in a related context by permitting the President to exempt military personnel accounts from a defense sequester provided the same total defense reduction was achieved and subject to congressional approval. That provision did not alter the equal split, and the condition attached to the provision confirmed the underlying rule. Relief for one defense account was permissible only when another defense resource absorbed the amount, because the category total remained fixed by the formula.

Annually appropriated programs formed the other group on which the burden settled with particular force. Appropriated activities depended on budget authority provided through the annual process, and those activities lacked the categorical exemptions that shielded the major benefit programs named in the statute. When a uniform percentage was applied within the nondefense category, the exposed portion consisted heavily of programs whose resources came through appropriations, together with any mandatory programs Congress had left unprotected. The benefit programs that dominated federal spending stood outside the calculation, the health programs in the special rule tier contributed at capped rates, and the appropriated remainder supplied the balance. Education provided a visible example across several years. The Education Department reduction under the fiscal 1986 sequester was $170.9 million. Office of Management and Budget figures during the fiscal 1988 confrontation implied an education reduction of about $1.4 billion. Under the fiscal 1990 sequester that took effect on October 16, 1989, discretionary programs were cut by 5.3 percent and the Education Department lost about $949 million to $1 billion on paper. Each figure described the experience of a single appropriated department inside the unprotected remainder.

The political logic that produced such a structure requires careful attribution. In one register, the exemption list recorded the boundaries of a legislative coalition, since programs with large beneficiary populations, programs delivering health care, and programs supporting veterans carried political weight that a deficit statute ignored at the cost of enactment. In another register, the exemptions recorded a substantive judgment that automatic cancellation was an inappropriate instrument for certain obligations, whatever the deficit figure in a given year. The National Affairs analysis of the act, published in Spring 2013, adopted the coalition reading in direct terms and described the exemptions as a concession that was the price of passage in the Democratic House. That attribution matters. The reading belongs to that analysis, offered with the benefit of the intervening record, rather than to a neutral description of the statutory text. The text itself listed programs and set caps without announcing a motive. The analysis connected the list to the legislative transaction that carried the Gramm-Rudman-Hollings amendment through a chamber where protection for the named programs formed part of the condition for agreement.

Read through that analysis, the exemption structure explained both why the act passed and why the act strained under operation. The concession reading held that Congress bought enactment by surrendering reach. A statute that exposed Social Security, Medicaid, veterans’ benefits, nutrition programs, and the other named accounts to automatic cancellation would have confronted opposition that the amendment, attached to a debt limit measure that had to pass, could ill afford. Protection disarmed that opposition program by program. Each exemption removed a constituency from the coalition against the bill, and each cap softened resistance from interests connected to the special rule health programs. The price appeared later, in the percentages. The same clauses that made the vote possible made the remedy severe for everyone outside the clauses, and severity created political pressure of another kind when an order threatened reductions of the size calculated for fiscal 1988 or fiscal 1991.

The administrative burden created by that concentration fell on agencies operating the unprotected programs, and the burden differed in kind from the ordinary task of managing a reduced appropriation. A conventional reduction enacted through legislation arrives with program specific choices attached, since Congress selects the accounts to reduce and can phase a change, protect a priority within an account, or adjust related authorities. A sequestration order arrived as a percentage. The order cancelled budgetary resources across accounts at the calculated rate, and the agency administering each account had to translate a uniform figure into operational decisions under constraints that the percentage did not recognize. Contracts already signed bound agencies to payments for work ordered or performed. Grant commitments, procurement schedules, and personnel obligations created fixed near term costs inside accounts whose totals the order reduced by a fraction calculated at the level of the category. Cancellation of a percentage of resources did not cancel a corresponding percentage of obligations, and agency managers faced the task of honoring commitments from a diminished pool while deciding which uncommitted activities would absorb a disproportionate share of the loss.

Forward funding deepened that difficulty and distorted the timing of the visible effects. Most education programs are forward funded, as the fiscal 1990 record noted, so a sequester applied to those programs produced a practical effect that lagged behind the order itself. Money for a school year or program cycle had often been structured to become available on a schedule set apart from the fiscal year in which the budget authority was provided, and a reduction recorded on paper in one period reached institutions in a later period. Planners could not match the reduction to the moment of announcement, and recipients could not adjust at the moment the order issued. An across the board order presented itself as a single dated event, effective on a stated day such as March 1, 1986 or October 16, 1989, while the administrative consequences unfolded across contract periods, grant cycles, and funding schedules that obeyed no single date.

The interaction between the exemptions and that burden produced the deepest strain inside the unprotected remainder. An agency account did not present itself to a sequestration calculation as a pool of equally flexible dollars. Part of an account might be committed under contract, part might support forward funded activity whose reduction would land in a later cycle, part might cover costs that could not be reduced without breaching an obligation, and part might remain genuinely available for adjustment in the year of the order. The uniform percentage applied to the account total. The flexible portion therefore faced an effective reduction larger than the stated percentage, because the inflexible portions contributed little or nothing in the near term while the total cancellation remained fixed. Exemptions at the level of the whole budget and rigidity at the level of the individual account operated by the same logic at different scales. Each layer of protection, statutory in one case and contractual or structural in the other, transferred weight to whatever remained movable. The structure also connected abstract percentages to departmental figures. A category rate of 4.3 percent or 5.3 percent described the formal calculation for a group of accounts, while a departmental figure such as $170.9 million described the same event at the level where contracts were managed and forward funded programs were planned. Removal of the fully exempt programs, caps on the special rule health programs, special calculation rules for guaranteed student loans and child support enforcement, and treatment of net interest as exempt in operation left the remaining accounts to generate the entire required reduction at whatever percentage the division produced.

Bowsher v. Synar and the constitutional defect

The constitutional weakness of the Balanced Budget and Emergency Deficit Control Act of 1985 stood at the center of the statute from the moment of enactment on December 12, 1985, because the mechanism that gave the law its force depended on a single assignment of power that the Constitution did not permit. Congress had fixed a descending schedule of maximum deficit amounts, running from $171.9 billion for fiscal 1986 to balance in fiscal 1991, and had provided that an estimate exceeding a target by more than a narrow margin would produce automatic cancellation of budgetary resources, divided equally between defense and nondefense categories and applied as a uniform percentage reduction inside each category. A schedule of numbers, however, could not cancel anything by itself. Some officer had to receive the estimates, evaluate them, calculate the required reduction, and report the determination that set the presidential order in motion. The act assigned that binding determination to the Comptroller General, the head of the General Accounting Office, the audit agency of the legislative branch. If the officer holding the decisive power held it unlawfully, the automatic process had no lawful engine, and that was the defect that Bowsher v. Synar exposed and resolved in less than eight months.

The sequence of reports written into the statute shows how completely the design routed executive action through the Comptroller General. The Office of Management and Budget and the Congressional Budget Office issued a joint deficit and sequestration report in August, before the fiscal year began, so the trigger rested on an estimate rather than on an outcome recorded after spending had occurred. The Comptroller General then issued a report on August 25 evaluating the joint submission, and the President issued an initial sequestration order on September 1, with the final order following on October 15. Updated reports from the Office of Management and Budget and the Congressional Budget Office arrived on October 5, a further report from the General Accounting Office followed on October 10, and a compliance report closed the cycle on November 15. The presidential order implemented a calculation that the Comptroller General had evaluated and settled, which meant that the officer who effectively decided whether sequestration occurred, and in what amount, sat outside the executive branch while performing a function that belonged to execution of the law.

The removal provision attached to the office turned that placement into a constitutional problem rather than a question of administrative convenience. The Comptroller General was removable by Congress, by joint resolution for inefficiency, neglect of duty, or malfeasance, or by impeachment. An officer who could be removed by the legislature for neglect of duty had reason to attend closely to the preferences of the legislature, and the statute placed that officer in charge of a decision with enormous consequences for programs that members of Congress had created, funded, and defended. Congress obtained the benefit of automatic reductions without recording a vote for the specific cancellations that followed, and Congress retained a removal lever over the officer whose report made those cancellations happen. The arrangement therefore combined distance and control in a single structure: distance from the political cost of particular reductions, and control over the official who announced them. That combination, rather than any minor error in a subordinate clause, formed the structural problem that litigation reached within hours of the signing ceremony.

One of the authors of the act, quoted in Congressional Research Service Report R41901, stated the purpose in terms that clarify why the design took the shape it did: “It was never the objective of Gramm-Rudman to trigger the sequester; the objective of Gramm-Rudman was to have the threat of the sequester force compromise and action.” A threat of that kind works only if the threatened outcome follows automatically once the stated conditions appear, because a threat that depends on a further discretionary decision invites negotiation over the decision itself and loses credibility as a forcing device. Automatic operation, in turn, required a decision maker positioned outside the ordinary bargaining between the White House and Capitol Hill, since either elected branch, given a free choice at the moment of decision, might prefer delay or exception to the pain of uniform reductions. Placing the determination in an office answerable to Congress did not remove the determination from the separation of powers; the placement relocated the determination to the one branch that the Constitution bars from executing the laws it enacts. The harder the statute worked to make the trigger automatic, the more decisive the report became, and the more clearly that report amounted to execution.

President Ronald Reagan identified the difficulty at the moment of signing, before any court had ruled. The signing statement recorded the view that the act was constitutionally defective because of the Comptroller General’s supervisory role, a position cited at 21 Weekly Comp. Pres. Doc. 1491 (1985). The statement nevertheless matters to the litigation history because the objection that prevailed in the Supreme Court was neither hidden nor improvised after the fact. The President who signed Public Law 99-177 on December 12, 1985 announced, on that occasion, the same structural objection that challengers carried into court: a supervisory role for an officer of the legislative branch over the execution of a fiscal statute crossed the boundary between making law and carrying law into effect. The first sequestration cycle therefore proceeded under a legal cloud that no participant could plausibly describe as unforeseen.

The challenge arrived with a speed that matched the warning. Congressman Mike Synar of Oklahoma, who had voted against the act, filed suit within hours of the signing, and eleven other members of Congress joined the action. The plaintiff was therefore not a distant regulated party discovering an injury after implementation, but a legislator who had lost the vote on the measure and who contended that the structure itself violated the Constitution. Twelve members of the House challenged a mechanism enacted by majorities of 271 to 154 in the House and 61 to 31 in the Senate, asking the judiciary to invalidate the reporting provisions on which automatic sequestration depended.

A parallel action by the National Treasury Employees Union supplied a different and more concrete grounding for the controversy. The union’s members were injured by the suspension of cost-of-living benefit increases, so the union’s suit rested on a direct financial effect felt by identifiable federal employees rather than on an institutional objection to the allocation of power among the branches. The two challenges complemented each other. The members of Congress presented the separation-of-powers question in its purest form, as a dispute about which branch could constitutionally hold the power to trigger automatic reductions. The union presented the monetary consequence of the mechanism in operation, the loss of scheduled increases for employees who depended on them.

The first judicial decision came from a three-judge district court on February 7, 1986, in Synar v. United States, 626 F. Supp. 1374 (D.D.C. 1986). The panel consisted of Judges Scalia, Johnson, and Gasch, and the court invalidated the reporting provisions of the act. The choice of provisions mattered. The court directed its ruling at the reporting structure that empowered the Comptroller General, the same assignment that the signing statement had questioned and that the Synar complaint had attacked, rather than at the deficit targets that Congress had enacted as policy. Invalidation of the reporting provisions struck the mechanism while leaving the stated fiscal goals without an automatic enforcement device, a distinction that the Supreme Court later preserved. The February date also mattered, because the fiscal 1986 cycle operated under a special rule: the sequester for that year was capped at $11.7 billion and the order took effect on March 1, 1986, on a seven-twelfths proration of the year. A ruling issued on February 7 therefore landed three weeks before the first cancellations took effect, close enough to disrupt the entire cycle if the judgment had immediate force, and the panel’s handling of its own ruling determined whether the first sequester would occur at all.

The district court stayed its own ruling pending Supreme Court review, and that stay allowed the March 1, 1986 sequester to take effect on schedule. The realized reductions under the capped fiscal 1986 order ran to approximately 4.3 percent for nondefense programs and 4.9 percent for defense programs, with the Education Department reduction recorded at $170.9 million in contemporaneous reporting. The episode produced an unusual legal posture: a federal court had held the reporting provisions invalid, yet the cancellations authorized through those provisions proceeded because the court suspended the effect of its judgment until the Supreme Court decided the question. For the litigants, the stay raised the stakes of the appeal, since affirmance would confirm that reductions of several billion dollars had been imposed under a defective procedure, while reversal would restore full legality to a process that had operated throughout the dispute.

Argument in the Supreme Court took place on April 23, 1986, and the decision followed on July 7, 1986, in Bowsher v. Synar, 478 U.S. 714 (1986). The full course from enactment on December 12, 1985 to a definitive constitutional judgment occupied less than eight months, an interval that included a three-judge district court proceeding, a stayed judgment, an operating sequester, and briefing and argument in the Supreme Court. The compressed schedule had substantive significance. The deficit target for fiscal 1987, $144.0 billion with a $10 billion margin, was approaching its own reporting cycle as the decision issued, so the ruling arrived in time to govern the second year of the statute rather than after the mechanism had entrenched itself across several cycles.

What exactly did the Supreme Court hold?

The Court held that the Comptroller General performed execution of the laws under the act, and that Congress could not assign execution to an officer removable by Congress. The automatic sequestration mechanism therefore fell, while the statutory deficit targets themselves remained in place.

The reasoning turned on the character of the functions assigned, and the majority treated that character as a question of substance rather than of labels. The Comptroller General’s duties under the statute were not limited to gathering information for Congress, auditing expenditures after the fact, or advising the President on the state of the accounts. The officer evaluated the joint report of the Office of Management and Budget and the Congressional Budget Office, determined the reduction required by the statutory formula when the estimated deficit exceeded the target and margin, and reported the determination on which the President issued the sequestration order. In the vocabulary of the decision, the Comptroller General’s functions under the act were the execution of the laws. Once the functions were classified as execution, the removal provision decided the case. An officer removable by Congress through joint resolution for inefficiency, neglect of duty, or malfeasance, or by impeachment, remained subject to congressional control in a manner incompatible with executive power, and Congress could not vest execution in an officer under that degree of legislative control. The defect therefore lay in the conjunction of two features, each lawful in other settings: Congress may create an audit office answerable to itself, and Congress may provide for automatic reductions by formula, but Congress may not assign the execution of the formula to its own removable officer.

The alignment of the justices gave the holding a breadth that a narrow majority would have lacked. Chief Justice Burger wrote the majority opinion, joined by Justices Brennan, Powell, Rehnquist, and O’Connor. Justice Stevens concurred, joined by Justice Marshall, bringing the judgment to seven votes. Justices White and Blackmun dissented. A judgment supported by seven justices on the central proposition, that execution of the laws cannot be vested in an officer removable by Congress, left little room for a later Congress to revive the same assignment in altered dress or for a later litigant to treat the ruling as provisional. The dissent preserved a record of disagreement, but the size of the majority settled the institutional question with a clarity that matched the speed of the proceedings: the reporting provisions could not be saved by minor adjustment, because the objection reached the identity and accountability of the officer who performed them.

Precision about what the Court struck, and what the Court left standing, is essential to understanding the aftermath. The automatic mechanism fell. The deficit targets themselves were not struck down. Congress remained free to pursue the schedule of maximum deficit amounts through ordinary legislation, through appropriations written to conform to the targets, through revenue and mandatory spending changes enacted by both chambers and presented to the President, or through a redesigned enforcement procedure that placed execution in constitutionally permitted hands. What Congress lost was the shortcut: the capacity to make reductions happen without further legislative action, on the authority of a report issued by an officer accountable to Congress. Enforcement reverted to politics, to negotiation between the branches, and to the expedited fallback procedure that the statute itself had supplied in anticipation of exactly the judicial outcome that occurred.

Section 274(f) of the act contained that fallback, and its presence shows that Congress legislated with the constitutional risk in view. Under the fallback procedure, a Temporary Joint Committee on Deficit Reduction, made up of members of the Budget Committees, would receive the report of the Office of Management and Budget and the Congressional Budget Office. The process would then lead to a joint resolution that Congress voted on under expedited rules. If the resolution passed and was signed, the resolution served as the basis for a presidential sequestration order. Expedited rules compressed debate and amendment, preserving some of the forcing effect that the sponsors had sought, but expedition is a matter of internal procedure, and internal procedure does not transfer execution to the legislature. The fallback thus traded automatic operation for constitutional safety: reductions could follow the reporting cycle on an accelerated timetable only after the elected branches had acted in their assigned capacities, Congress by voting and the President by signing and ordering.

The comparison between the invalidated mechanism and the section 274(f) fallback clarifies why the defect cannot be dismissed as a technicality about personnel. A technical defect can be cured by relabeling an office or adjusting a reporting date, leaving the underlying allocation of power intact. The Bowsher defect required the opposite remedy: the allocation of power itself had to change, because the officer who made the binding determination had to be removed from the chain of execution, and the determination had to be replaced by action from the branches that the Constitution authorizes to legislate and to execute. The fallback accepted delay, debate, and the possibility of defeat on the floor in place of certainty, and that acceptance amounted to a concession that the automatic feature, the feature that made the statute distinctive, depended on the unconstitutional assignment. The whole scheme had been built to take the decision away from the elected branches at the moment when reductions became necessary, on the theory that elected officials facing that moment would find reasons to postpone it. A decision removed from the elected branches had to be lodged somewhere, and the statute lodged the decision in an officer answerable to Congress, which the Constitution forbids for execution. The design therefore contained a contradiction that no refinement of wording could dissolve: credibility required automaticity, automaticity required an outside decision maker, and the outside decision maker chosen was the one category of officer disqualified from executing the law.

That contradiction also explains why the litigation moved so quickly and why the outcome commanded such a broad majority. Challengers did not need to contest estimates, dispute the fiscal record, or argue that particular reductions were unwise. The objection operated at the level of structure, available on the face of the statute from the day of signing, and capable of resolution without a factual record about program effects. Congressman Synar filed within hours because nothing further needed to happen before the defect could be presented to a court, and the National Treasury Employees Union could point to suspended cost-of-living benefit increases as the concrete injury that the mechanism produced. The three-judge court ruled within two months, the Supreme Court heard argument in April, and the final decision issued in July, all within eight months of enactment. The decision closed that chapter by removing the automatic trigger while leaving the targets, the fallback procedure, and the ordinary legislative process in place, and the repair legislation of 1987 later rebuilt the trigger around the Director of the Office of Management and Budget, an executive officer, in the Balanced Budget and Emergency Deficit Control Reaffirmation Act of 1987, Public Law 100-119. The sequence from signing statement to judgment to redesign confirms the place of Bowsher v. Synar in the history of the act: the case did not interrupt an otherwise sound design from outside, but identified the flaw that the design had carried from its first day, in the assignment that made automatic government by formula appear possible.

The repair and the operating record

The judgment in Bowsher v. Synar left Congress with targets on the books and no lawful automatic device for enforcing them. The Court held that the functions assigned to the Comptroller General under the 1985 act amounted to execution of the laws, and that an officer removable by Congress could not be vested with that execution. The deficit ceilings themselves were not struck down. What fell was the machinery that had converted an estimate into a presidential order without a further vote. A legislature that wished to preserve the discipline therefore had to rebuild the trigger around an officer inside the executive branch, and it had to do so while the fiscal calendar continued to run under targets written for a mechanism that no longer operated in its original form. A measure styled a Reaffirmation Act had been signed on August 21, 1986, after the decision, but the operative repair arrived the following year. The repair that restored the process took the form of the Balanced Budget and Emergency Deficit Control Reaffirmation Act of 1987, Public Law 100-119, enacted as H.J.Res. 324 and signed by President Ronald Reagan on September 29, 1987.

The vehicle repeated the pattern of 1985. Just as the original act had traveled on a measure raising the statutory debt limit, the 1987 repair was again carried on a debt limit increase, and the limit enacted in that measure stood at $2.8 trillion. The pairing mattered for the legislative history as well as for the calendar. A debt limit bill had to move, because the government required the borrowing authority, and a measure that had to move offered a route for restoring a budget process that might otherwise have stalled amid disagreement over whose estimate should control and how much additional time the targets should allow. Congress used that route to do more than substitute one reporting officer for another. The 1987 act rewrote the schedule, adjusted the conditions under which an order could be displaced, and added a limited flexibility for defense accounts, all while preserving the basic architecture of an estimated shortfall, a determination, and a uniform percentage cancellation of budgetary resources divided between defense and nondefense categories.

What changed when the determination moved to the budget office?

The Director of the Office of Management and Budget became the official whose report triggered sequestration for fiscal 1988 through fiscal 1993. The Congressional Budget Office reported to the budget office and to Congress, and the budget office had to follow its technical assumptions and methodologies and explain any discrepancy between the two reports.

That allocation of roles answered the constitutional defect without removing Congress from the calculation altogether. The Congressional Budget Office, an agency of the legislature, retained a reporting function, and its report went both to the Office of Management and Budget and to Congress itself. The executive officer who issued the operative determination could not treat the legislative estimate as an unrelated opinion. The statute bound the Director to the technical assumptions and methodologies of the Congressional Budget Office, and it required an explanation for any discrepancy between the two reports. A difference in the bottom line therefore had to be traced and defended in the open, rather than absorbed silently into a single executive figure. The design preserved a check on executive optimism at the level of method, which was the level at which the earlier disputes over economic assumptions had done their work. It did not give the Congressional Budget Office the last word on the determination. It gave that office a documented benchmark against which the determination could be read, and it made departures from that benchmark a matter of record for the committees that would decide whether to accept, displace, or legislate around an impending order.

The second major change concerned time. Under the 1985 schedule, balance was to be reached in fiscal 1991, with the final step carrying the target from $36.0 billion in fiscal 1990 to zero. By 1987 that endpoint had ceased to describe a plausible path, because the intervening years had not produced the outcomes the earlier schedule assumed. The Reaffirmation Act moved the balance date from fiscal 1991 to fiscal 1993 and rewrote every intervening ceiling. The revised target for fiscal 1988 was $144.0 billion, and for that year the statute allowed no margin, so an estimate above $144.0 billion exposed the government to sequestration. The target for fiscal 1989 was $136.0 billion, with a margin of $10 billion. The target for fiscal 1990 was $100.0 billion, with a margin of $10 billion. The target for fiscal 1991 was $64.0 billion, with a margin of $10 billion. The target for fiscal 1992 was $28.0 billion, with a margin of $10 billion. Fiscal 1993 carried a target of zero, and no margin applied at that endpoint. The sequence therefore paired a two year extension with a near term reset, since the fiscal 1988 ceiling of $144.0 billion repeated the nominal figure that the original schedule had set for fiscal 1987, and the reductions demanded in the middle years returned to the pattern of large annual steps toward balance.

The margin deserves attention because it defined the real trigger in most years. A target stated the maximum amount of deficit the law professed to allow, while the margin stated the amount by which an estimate could exceed that maximum amount before cancellation followed. For fiscal 1989 through fiscal 1992, a $10 billion margin meant that the operative line for enforcement stood $10 billion above the printed ceiling. The endpoints were treated differently. Fiscal 1988, the first year under the restored mechanism, had no margin, and fiscal 1993, the year of promised balance, had no margin either. Between those endpoints the law conceded a band of estimation tolerance, reflecting the fact that the trigger ran on a forecast made before the fiscal year began rather than on an audited result after its close. That concession also shaped behavior, because a package that brought an estimate inside the target plus the margin satisfied the statute even when the same package left the printed target unmet, and later results could diverge from the estimate without reopening an order for a year that had already passed its determination point.

Two further features of the 1987 act altered the consequences of an order without abandoning the equal division at its center. The President received an option to exempt military personnel accounts from a defense sequester, provided that the same total defense reduction was achieved elsewhere and subject to congressional approval. The option recognized that a uniform percentage cut applied across defense accounts could fall on personnel costs in ways that disrupted force structure, yet the condition attached to it prevented the exemption from becoming a reduction in the defense contribution itself. The savings had to be found within the defense category, and Congress retained an approval role over the choice. Separately, the act created an expedited joint resolution procedure by which Congress could supplant a final sequestration order. That route gave the legislature a defined, fast means of substituting its own enacted judgment for the automatic allocation, instead of leaving members to choose between accepting an order in full and assembling an ordinary bill under ordinary timing. Both features loosened the rigidity of the 1985 design at points where rigidity had threatened either military administration or legislative responsibility, while leaving the central bargain intact: a missed estimate produced cuts, divided between the two categories, unless Congress legislated a different result.

The record that followed is best read year by year, because the mechanism behaved differently in each fiscal cycle and because the distinction between an estimate and an outcome sits at the center of every year in the sequence. The figures for outcomes in this section are the unified deficit in the Office of Management and Budget Historical Tables, cited by fiscal year, and they measure what the government recorded after each year closed. The targets and trigger levels, by contrast, governed decisions taken in advance, on estimates assembled before or near the start of a fiscal year. A year could therefore satisfy the law and miss its target, or face cancellation pressure and then record an outcome shaped mainly by legislation negotiated under that pressure. The five years from fiscal 1986 through fiscal 1990 supplied examples of each combination, and together they explain both the influence the statute exercised over congressional behavior and the weakness that later led legislators toward a different form of enforcement.

Fiscal 1986 operated under a special rule written for a year already under way when the act became law in December 1985. The sequester for that year was capped at $11.7 billion, and the order took effect on March 1, 1986, after the three judge district court had invalidated the reporting provisions and stayed its own ruling pending Supreme Court review. The capped amount limited the immediate disruption, and a seven twelfths proration recognized that much of the fiscal year had already elapsed before the cancellation reached agency accounts. The outcome, however, bore little relation to the ceiling around which the year had been organized. The fiscal 1986 unified deficit in the Office of Management and Budget Historical Tables was $221.2 billion, against a target of $171.9 billion drawn from the budget resolution assumption for that year. The miss carried an additional lesson about forecasting. At passage, analysts had projected a fiscal 1986 deficit of roughly $180 billion, a figure far closer to the target than to the recorded result. The recorded result exceeded that projection by a wide amount, and the law had embedded the more hopeful path in its schedule. An enforcement system built on an estimate inherited every error in that estimate, and fiscal 1986 demonstrated the scale of error possible within a single year, before the Supreme Court had even delivered its judgment on the mechanism itself.

Fiscal 1987 produced the single year in which the process cleared its own test. The joint estimate prepared under the original procedure came in at or within the target plus the margin, with a target of $144.0 billion and a trigger level of $154.0 billion, so no sequester was ordered for that year. On the terms the statute set for itself, the year was a success: the advance calculation did not cross the line, the automatic device stayed sheathed, and the government proceeded without a cancellation order. The qualification attached to that success is essential to understanding the design. The fiscal 1987 unified deficit in the Office of Management and Budget Historical Tables was $149.7 billion, which exceeded the $144.0 billion target even though it fell inside the band the margin had created around the estimate. The clean year was clean on an estimate, not on a result. Nothing in the procedure returned to the year after the accounts closed and demanded a correction for the difference, because the law had chosen the snapshot as its moment of judgment. Supporters of the design could point to a year without disruption and a recorded deficit lower than the year before, while the arithmetic of the target itself recorded a miss, and both readings rested on figures the statute treated as answering different questions.

Fiscal 1988, the first year governed by the repaired mechanism and the revised schedule, turned on events outside the budget process altogether. After the stock market crash of October 19, 1987, a threatened sequester of about $23 billion framed negotiations between the White House and Congress. The threatened cancellation was split evenly, with $11.5 billion assigned to defense and $11.5 billion to nondefense, and the Congressional Budget Office calculated the uniform reduction percentages required within each category at 10.4 percent for defense and 8.7 percent for nondefense. Percentages of that size, applied across accounts that included programs with fixed obligations and accounts already operating under enacted appropriations, concentrated attention in a way that another round of resolutions had not. A sequestration order was issued in November 1987, and the scheduled cuts stood as the alternative against which a negotiated package would be measured. The budget summit that followed produced an agreement under which the fiscal 1988 deficit would be reduced by about $30.2 billion, to roughly $149.7 billion on congressional estimates, with nearly $46 billion of reduction for fiscal 1989. The two year package was described as $76 billion of deficit reduction, a figure that gave both branches a settlement larger than the single year cancellation it displaced.

The summit agreement required legislation, and Congress supplied it before the end of the calendar year. The Omnibus Budget Reconciliation Act of 1987, Public Law 100-203, signed on December 22, 1987, carried the revenue and mandatory spending elements of the settlement, and an omnibus appropriations act, Public Law 100-202, enacted in the same period carried the appropriations elements. Once the agreement had been enacted, the November sequestration order was canceled, and the scheduled reductions did not proceed on their original terms. The sequence illustrated the purpose that one of the act’s authors had described for the device, since the threat had forced a compromise that produced a legislated package exceeding the threatened cancellation in stated size. The outcome for the year told a more restrained story. The fiscal 1988 unified deficit in the Office of Management and Budget Historical Tables was $155.2 billion, against the revised target of $144.0 billion for that year, a target that had allowed no margin. Congress had acted, the automatic cuts had been displaced by a negotiated settlement, and the recorded deficit nevertheless stood above the ceiling the repair had set only months earlier, a pattern that separated legislative compliance with a summit agreement from attainment of the statutory number.

Fiscal 1989 tested whether a settlement negotiated in one autumn could carry a second year without a fresh crisis. The summit package had assigned nearly $46 billion of its reduction to fiscal 1989, and that second year of the agreement carried the year through without a sequester. No order dominated the calendar, and the machinery receded from view precisely because the earlier bargain had already allocated the savings the year required under its estimates. The recorded result repeated the pattern of the year before in quieter form. The fiscal 1989 unified deficit in the Office of Management and Budget Historical Tables was $152.6 billion, against a revised target of $136.0 billion. The gap between result and ceiling widened in a year with no crash, no threatened order of the size faced in fiscal 1988, and no dramatic confrontation to explain the divergence. That widening mattered for the credibility of the revised schedule, because fiscal 1990 required a step down to a $100.0 billion target, and a government recording deficits above $150 billion faced that step with little evidence that ordinary adherence to a prior agreement would close the distance. The margin of $10 billion softened the trigger, yet even the trigger level stood far below the outcomes the government had recorded in two consecutive years under the repair.

Fiscal 1990 supplied the year in which the automatic device actually took effect and remained, in reduced form, part of the final settlement. Lawmakers missed the October deadline for completing action that would have displaced an order, and a sequestration order issued on October 16, 1989, cut discretionary programs by 5.3 percent. The cancellation reached across the nondefense discretionary base, and the Education Department recorded a reduction of about $949 million on paper. The practical effect in education arrived more slowly than the paper figure suggested, because most education programs were forward funded, meaning that money appropriated in one fiscal year financed school activity in a later period. A reduction entered in the accounts for fiscal 1990 therefore pressed on classrooms with a delay built into the funding structure, even though the legal cancellation had already occurred. That lag complicated the politics of the order in both directions. The cuts were real in budgetary terms from the date of the order, while the visible consequences in services were postponed, which weakened the immediate pressure that an in year cancellation was supposed to exert and complicated any later claim that restoring the funds would restore the service on the original timetable.

Congress answered the October order with reconciliation legislation rather than with a full restoration. The Omnibus Budget Reconciliation Act of 1989, Public Law 101-239, signed on December 19, 1989, carried an estimated $14.7 billion of deficit reduction, about half of the summit agreement figure associated with that year in the reporting on the act. Title XI of that statute addressed the sequester directly. It rescinded the October 16, 1989 order and directed a new order carrying adjusted and smaller reductions, so the final cancellation for fiscal 1990 was lower than the cancellation that had taken effect in October. Title XI also prohibited any double reduction in Medicare, protecting that program against being cut twice through the interaction of the original order and its replacement. The same title legislated the treatment of transfers across the fiscal year boundary, providing that transfers of outlays or receipts under the act would be treated as a necessary result of a significant policy change under the 1987 act, so that such transfers counted for purposes of the budget process. That provision confirmed, in statutory text, that timing shifts across the September 30 boundary had become part of the practical contest over the targets. The partial sequester largely survived in reduced form, since Congress kept part of the automatic cuts rather than restoring the affected accounts in full, and the year closed with a settlement in which legislated reductions and a scaled back cancellation operated together.

The fiscal 1990 outcome closed the period on the widest divergence in the sequence. The fiscal 1990 unified deficit in the Office of Management and Budget Historical Tables was $221.0 billion, against the revised target of $100.0 billion set by the 1987 repair. The recorded figure was more than double the ceiling, and it arrived in a year in which the mechanism had functioned in its most complete form: an estimate had triggered an order, the order had taken effect, Congress had legislated in response, and the replacement order had preserved a portion of the automatic reduction. Procedure, in other words, had run its full course without bringing the outcome near the number the procedure existed to reach. Read against the earlier years, the pattern was consistent in direction and varied only in scale. Fiscal 1986 missed a target shaped by an optimistic forecast. Fiscal 1987 satisfied the trigger and missed the target. Fiscal 1988 displaced an order through a summit and missed the revised target. Fiscal 1989 relied on the second year of that summit and missed by a larger amount. Fiscal 1990 let the device operate and missed by the largest amount of the five years. The repair of 1987 had restored a lawful trigger, bound the executive determination to legislative methods, extended the road to balance by two years, and added measured flexibility on military personnel accounts and on congressional displacement of an order. The operating record through fiscal 1990 showed the limits of each of those achievements, because a law that judged years in advance, on estimates, against ceilings that legislatures could rewrite, could shape negotiations powerfully while leaving the recorded deficit to forces and decisions that the trigger never directly governed.

Evasion by mechanics: satisfying a projection on paper

The techniques described in this section shared a common structural starting point. The statute did not test a completed fiscal year against its target after the receipts had been collected and the outlays had been recorded. The test ran earlier, on an estimate prepared before the fiscal year began. The Office of Management and Budget and the Congressional Budget Office issued a joint deficit and sequestration report in August, the Comptroller General issued an evaluation on August 25, and the presidential sequestration order followed in September and October. A determination made at that distance from the outcome depended on assumptions about economic growth, revenue, spending rates, and the timing of payments, and each of those elements could change without any change in the underlying programs themselves. The margin, the scoring conventions, and the calendar all entered the calculation before a dollar for the year in question had been spent, so the calculation itself became the field on which compliance was established.

That structure placed unusual weight on a single projected number. If the August estimate placed the deficit at or below the target plus the permitted margin, no sequestration order followed, regardless of the figure that the completed year eventually produced. If the estimate placed the deficit above that line, the automatic cancellation of budgetary resources followed unless Congress and the President enacted reductions of their own. Because the consequence turned on the estimate, every input to the estimate acquired practical importance equal to an enacted reduction, and none of those inputs required a program to operate at a lower cost over time. The techniques differed in form, but each operated on the estimate rather than on the longer run cost of the government.

Optimistic economic assumptions formed the first and most direct technique, because an assumption required no legislation, no sale, and no change in a payment schedule. A budget plan built on stronger projected growth showed higher projected receipts, since a larger economy generated more taxable income in the estimating models. The same plan showed lower projected outlays for programs whose costs moved with economic conditions. The combined effect reduced the projected deficit on paper at the moment of the estimate, and the statute tested the plan at exactly that moment. The sensitivity ran in both directions. At passage, analysts had forecast roughly $180 billion for fiscal 1986, and the completed year produced a figure more than $50 billion above the forecast path embedded in the law. Forecast error of that size meant that the choice among forecasts determined whether the trigger fired.

The fiscal 1988 episode gave the technique a documented form. The Reagan administration presented a fiscal 1988 budget plan that complied with the $108 billion target then in effect, and the compliance rested on economic forecasts that supported the projected result. Senator Lawton Chiles of Florida, the Chairman of the Senate Budget Committee, and Representative William H. Gray of Pennsylvania, the Chairman of the House Budget Committee, called the forecasts unrealistic. Their objection went to the inputs rather than to the arithmetic. The arithmetic, given the assumptions, produced compliance. Their argument was that less favorable assumptions produced a different result, and that the alternative result described the fiscal position more accurately. The disagreement therefore concerned which projection the statute would treat as decisive, and the administration plan held the advantage because its projection supported the conclusion that no further action was required for that year.

The Congressional Budget Office supplied the competing estimate. CBO estimated that the fiscal 1988 deficit would approach or exceed $170 billion, a figure about $20 billion above the administration forecast. A gap of $20 billion between two estimates for the same year measured the room that assumption choices created inside the process. The target for the year was $108 billion, so neither estimate described a year close to the statutory goal, but the practical question under the trigger was narrower than the distance from the target. The practical question concerned the size of the reduction that the estimate required. A lower forecast reduced the required action, even where the forecast left the projected deficit far above the target. Assumption choices thus influenced the process at two levels, first by determining whether the trigger fired at all in a year close to the line, and second by determining the scale of the response in a year far from the line.

The administration forecast also had to be read against a baseline that CBO had published earlier in the same cycle. CBO’s January 1987 baseline showed fiscal 1987 running near $174 billion against a $144 billion target for that year. Fiscal 1987 eventually closed with the fiscal 1987 unified deficit in the Office of Management and Budget Historical Tables recorded at $149.7 billion, below the January baseline figure and above the $144 billion target. That sequence showed the range within which estimates moved during a single year. Movement of that size, applied to the following year’s plan, could carry a plan across a compliance line without any enacted change in law.

How could a target be met without an equivalent real reduction?

A target could be met on the scored estimate because the trigger compared a forecast, not an outcome. Changing an assumption, selling an asset, moving a payment date, or placing spending outside the measured total altered the August projection enough to stay within the target plus margin without reducing spending by the same amount.

The margin supplied a second technique that required no action at all beyond the design of the statute itself. For fiscal 1987 through fiscal 1990, the estimated deficit could exceed the target by $10 billion without triggering a sequester. The margin had a mechanical purpose in a system run on estimates, since an estimate carries error and a trigger set exactly at the target would fire on small forecasting differences. In operation, however, the cushion changed the effective target for each year in which the margin applied. A plan that produced an estimate $10 billion above the stated target satisfied the test, and planners working toward the test worked toward the higher figure. An analysis published in National Affairs in 2013 described the result directly, concluding that the $10 billion margin functioned in practice as a standing upward revision of each target. The stated target announced the goal, the margin defined the enforceable line, and the distance between the two figures was available in every year without negotiation or legislation.

The margin also interacted with the estimating techniques, because the cushion absorbed the first portion of any unfavorable revision. If an estimate rose by less than the margin above the target, the rise had no consequence under the trigger. If an assumption change lowered an estimate toward the target, the margin completed the distance and converted a near approach into a pass. The fiscal 1987 year illustrated the combined operation. The estimate from the Office of Management and Budget and the Congressional Budget Office came in at or within the target plus the $10 billion margin, with a target of $144 billion and a trigger level of $154 billion, and no sequester was ordered. The completed year produced a deficit above the $144 billion target. The process had cleared its test on the estimate, within the margin, in the one year in which no sequester followed under the original schedule.

Asset sales formed a third technique, and the August snapshot basis gave the technique its effect. A sale of a government asset produced receipts in the year of the sale, and the snapshot counted those receipts against the deficit for that year. The transaction reduced the measured deficit for the target year even though the government surrendered an asset of value in exchange for the receipts, and even though the receipts arrived once while the programs charged against the target recurred each year. The sale of Conrail served as the standard example, with the fiscal 1987 Conrail sale used to generate paper savings against the targets, as the 2013 National Affairs analysis described. Sales of government loan assets served a similar function. A portfolio of loans represented future repayments flowing to the government over a period of years, and a sale converted that stream into a single payment in the year of the sale. The target year recorded the full amount, and later years lost the repayments that the portfolio would have produced, but the statute tested one year at a time and the snapshot captured the benefit at the moment of the test.

The scoring of loan sales carried a sensitivity that the Congressional Budget Office identified in its discussion of scoring conventions. Under the conventions CBO described, sales of loans with recourse were treated as borrowing rather than as offsetting collections. The distinction mattered because recourse left the government exposed to loss if the underlying loans failed to perform, so the transaction resembled a loan secured by the portfolio rather than a clean transfer of the asset. The contrast between the two treatments showed how much depended on the classification assigned to the transaction. A change in classification changed the scored result without any change in the cash that moved or the risk that the government retained. Asset sales therefore operated through two channels at once, the timing channel, which concentrated receipts in the target year, and the classification channel, which determined whether the receipts counted against the deficit at all.

Payment date shifting operated through timing alone. Moving a payday or a payment date by a day moved outlays between fiscal years, because an outlay belongs to the fiscal year in which the payment occurs. A payment scheduled for the first day of a new fiscal year and paid on the last day of the old year raised the old year’s total and lowered the new year’s total by the same amount. A payment moved in the opposite direction lowered the total for the year that a target test measured. The amounts involved in federal paydays and major payment cycles were large enough for a one day shift to move a measured deficit by a sum that mattered against a target and a $10 billion margin. The programs themselves cost the same across the two years combined, the recipients received the same payments with a difference of a day, and the only change with legal significance under the statute was the year to which the outlay was assigned. The snapshot structure rewarded exactly that assignment decision, since the test read the total for one year in isolation from the adjoining years.

The practice received express legislative treatment, which confirmed that cross year movement of payments formed part of the operating record. The Omnibus Budget Reconciliation Act of 1989, Public Law 101-239, signed on December 19, 1989, addressed the fiscal 1990 sequestration order in Title XI and dealt with transfers across the fiscal year boundary in the same title. Under that provision, transfers of outlays or receipts across the fiscal year boundary under the act were treated as a necessary result of a significant policy change under the 1987 act, with the consequence that the transfers counted. The provision placed the transfers inside the recognized arithmetic of the process. The same act rescinded the October 16, 1989 sequestration order in part and directed a new order with reduced reductions, so the statute that adjusted the sequester also settled the treatment of the timing transfers that accompanied the year’s bargaining.

A further group of techniques worked on the boundary of the measured total rather than on the total itself. Spending placed outside the budget, or designated under an emergency or disaster heading, or moved among appropriations accounts, changed the figure that the target test read without necessarily changing the amount the government spent. Shifting funding around within appropriations allowed an account under pressure from a reduction to draw on arrangements elsewhere in the appropriations structure, so the reduction recorded in one place did not translate into an equal reduction in activity. Emergency and disaster declarations supplied a route by which spending could proceed under a designation that carried different treatment from ordinary appropriations. Off budget placement removed an activity from the measured totals altogether. Each variant depended on a classification decision.

Mike Solon, a longtime adviser to Senator Gramm, described that group of practices in reporting published in 2013. Solon said, “They would shift funding around in appropriations, use emergency declarations, disaster declarations… All these sort of things to game it.” The description came from an adviser associated with one of the statute’s own sponsors, and the practices named in the quotation corresponded to the boundary techniques that the structure made available. An appropriations shift exploited the difference between an account and the whole. An emergency declaration exploited the difference between designated spending and ordinary spending. In each case the government’s total activity could proceed at a level close to the level planned before the test, while the measured figure presented to the trigger reflected the classifications rather than the total.

The final technique required no assumption, classification, sale, or calendar decision, because the technique amended the target itself. When the demands of the schedule could no longer be met through adjustments to the estimate, Congress revised the schedule. The Balanced Budget and Emergency Deficit Control Reaffirmation Act of 1987, Public Law 100-119, signed on September 29, 1987, raised the fiscal 1988 target from $108 billion to $144 billion, set a new sequence of targets running to a balance date in fiscal 1993 rather than fiscal 1991, and restored the mechanism under the authority of the Director of the Office of Management and Budget after the Supreme Court decision in Bowsher v. Synar, 478 U.S. 714 (1986). Revision solved the immediate compliance problem by redefining the standard against which compliance was measured. A deficit that failed the original target could satisfy the revised target without any change in revenue or spending, because the line had moved to the figure.

The sequence ended in 1990 with abandonment rather than another revision. The Omnibus Budget Reconciliation Act of 1990, Public Law 101-508, carried the Budget Enforcement Act of 1990 in Title XIII, and that title replaced the fixed deficit targets with discretionary spending caps and PAYGO enforcement over legislated changes to revenue and mandatory spending law. The revised targets for fiscal 1992 and fiscal 1993 never operated, and the balance date in fiscal 1993 arrived with a deficit of $255.1 billion in the unified series. The targets were superseded by the new enforcement design rather than formally repealed, and the distinction preserved the form of the original statute while its operative test ceased to govern. Amendment had first moved the targets, and replacement then removed the need to move them again.

David Primo of the University of Rochester, quoted in the 2013 National Affairs analysis, placed that ending inside a general pattern for automatic cutting mechanisms. In the pattern Primo described, when obedience to a mechanism became too costly, the mechanism itself was changed. Assumption changes, margin use, asset sales, timing shifts, and classification decisions each carried a limited capacity, measured by the size of the gap each could close. When the gap between the projected deficit and the target grew beyond the combined capacity of those techniques, the remaining option faced no capacity limit, since a legislature that sets a target by statute can reset the target by statute. The cost of that option was visibility rather than feasibility, because a revision announced in statutory text that the standard had changed.

Senator Ernest F. Hollings of South Carolina, the Democratic sponsor whose name stood third in the common name of the statute, delivered a personal verdict on the record in reporting published in 2013. As reported in that year, Hollings went to the Senate floor and said, “I want a divorce from Gramm-Rudman-Hollings. Y’all just call it Gramm-Rudman.” The remark separated the sponsor from the statute in the terms of the statute’s own title, and the request to remove one name from the common name recorded the distance between the design as enacted in 1985 and the operating record that followed. A sponsor’s request for removal of the sponsor’s name marked the point at which the gap between the announced schedule and the measured results became part of the public identity of the law itself.

Each technique in the catalogue exploited the same gap, between a legislated action and a projected outcome. A legislated action has definite terms that scoring can follow exactly. A statute that raises a tax rate, reduces a benefit formula, or cancels an appropriation produces a change that estimators can trace to text, and the successor framework built in 1990 chose that ground for enforcement, applying caps and PAYGO rules to legislated changes rather than to the deficit itself. A projected outcome has no equivalent firmness. An outcome depends on an economy that estimators model rather than observe, on payments that administrators schedule, on assets that officials may sell or hold, and on classifications that determine which totals the test reads. The outcome can therefore be re-estimated through an assumption, re-timed through a calendar decision, and re-defined through a classification or an amendment, and each operation changes the measured result while leaving the underlying fiscal position closer to its previous course than the measurement suggests.

The endgame, the successor design, and the lesson

The fiscal 1991 test exposed the central weakness of the Gramm-Rudman-Hollings design in a form that neither Congress nor the President could treat as an ordinary procedural problem. Under the revised schedule enacted in 1987, the maximum amount for fiscal 1991 was $64.0 billion, and the margin attached to that amount was $10 billion. A sequester therefore followed if the estimate placed the fiscal 1991 deficit above $74 billion. The test did not wait for an outcome recorded after the year closed. The statute judged a projection prepared before the year had run its course, and the projection for fiscal 1991 moved far beyond the trigger level that the automatic remedy lost any plausible claim to be a remedy at all. The distance measured the gap between a schedule written in 1987 and an economy and a federal budget that had refused to follow that schedule through the intervening years.

The Office of Management and Budget delivered its final sequestration report in October 1990 with an estimated deficit of $147.3 billion for fiscal 1991. Measured against the $74 billion trigger level, the estimate stood at nearly twice the figure that would have avoided automatic cancellation of budgetary resources. Measured against the $64.0 billion target itself, the estimate exceeded the permitted amount by $83.3 billion. A reduction program built from that estimate required defense outlay reductions of $41.7 billion, calculated at 34.7 percent, and nondefense reductions of $41.7 billion, calculated at 31.6 percent. The equal dollar division between the two categories reflected the architecture of the original mechanism, under which the total required reduction was split evenly between defense and nondefense. The percentages differed because the bases to which the equal dollar amounts applied differed, and the scale of each percentage conveyed what the equal division concealed. A cancellation approaching one third of the affected defense base and approaching one third of the affected nondefense base was a fiscal event of a different order from the smaller sequesters and threatened sequesters that had framed negotiation in earlier years.

No branch of the federal government intended to execute reductions of that size through that instrument. The executive branch had produced the report that described the reductions, and the legislative branch confronted the consequences of a law that the legislative branch had enacted and had repaired in 1987, and neither branch treated the October figures as an operating plan for fiscal 1991. A mechanism designed to discipline choice by threatening an intolerable outcome depends for its force on the credibility of the threat, and the fiscal 1991 figures tested credibility at the extreme. If the threat was executed, programs in both categories absorbed cancellations that the political branches had never chosen in that form. If the threat was avoided by legislation, the avoidance itself confirmed that the schedule yielded whenever the schedule demanded more than the political branches would accept. The statute had reached a point at which compliance and evasion both damaged the premise on which the law rested, because the premise had been that an automatic consequence could hold a numerical outcome in place regardless of the conditions that produced the outcome.

The response to the October estimate followed a path that earlier confrontations under the statute had already marked out. Political leaders repaired to a budget summit at Andrews Air Force Base, and negotiation at that summit produced a five-year agreement for $500 billion of deficit reduction. The summit format mattered to the meaning of the endgame. A law intended to replace discretionary bargaining with a fixed schedule and an automatic sanction ended its effective life in a bargained settlement reached away from the ordinary floor process, with the size and duration of the settlement set by negotiation rather than by the statutory table. The five-year horizon of the agreement also mattered. The Gramm-Rudman-Hollings schedule had proceeded year by year toward a single balance date, with each annual maximum amount derived from a straight descent to zero. The Andrews agreement addressed a longer span with a package, and a package can combine revenue decisions, mandatory spending decisions, and appropriations decisions in proportions that an across-the-board cancellation formula cannot select. The negotiators at Andrews chose composition, and composition had been exactly what sequestration was designed to avoid choosing.

Implementation came through the Omnibus Budget Reconciliation Act of 1990, cited as H.R. 5835 and enacted as P.L. 101-508, signed on November 5, 1990 by President George H. W. Bush. Title XIII of that act bore the name Budget Enforcement Act of 1990, and Title XIII did more than adjust the numbers in the predecessor statute. Title XIII replaced the fixed targets with a different enforcement architecture built on discretionary spending caps and on pay-as-you-go enforcement, called PAYGO, applied to legislated changes in revenue and mandatory spending law. The replacement was structural rather than cosmetic. A cap limits a category of spending that Congress enacts through appropriations. PAYGO scores the effect of bills that Congress passes and the President signs. Both instruments attach to decisions that identifiable officials take at identifiable moments, and a sequester under the successor design fell, when triggered, as a consequence of those scored decisions rather than as a judgment on the condition of the economy in a given year.

The shift from outcomes to legislated actions stands as the defining feature of the 1990 settlement, and the distinction deserves precise statement. Under Gramm-Rudman-Hollings, the legally consequential fact was an estimated outcome, namely the size of the shortfall for a fiscal year, a figure that depended on economic growth, on revenue collections that followed that growth, on interest costs, and on spending that earlier laws had set in motion. Under the Budget Enforcement Act of 1990, the legally consequential facts were legislated actions, namely appropriations that exceeded a cap or enacted changes whose scored cost lacked an offset. A legislature controls the second class of facts in a direct sense that does not hold for the first class. Members vote on an appropriations bill. Members vote on a tax provision or a benefit provision. No vote fixes the growth rate of the economy, and no vote fixes the revenue yield of a given tax structure under economic conditions that differ from the conditions assumed when the provision passed. The successor design therefore narrowed the object of enforcement to the object that enforcement could actually reach, and the narrowing amounted to a concession about the limits of procedural law that the 1985 and 1987 statutes had refused to make.

Precision also matters in describing what happened to the predecessor law in November 1990. The fixed targets were abandoned, and the statute that had carried those targets was superseded rather than repealed. The distinction between supersession and repeal is a recurring source of error in accounts of the period, and the record supports the narrower description. Congress did not enact a simple repeal that erased the Gramm-Rudman-Hollings framework as though the framework had never governed. Congress enacted a new enforcement design through Title XIII of P.L. 101-508, and the new design displaced the old one as the operative system of budget control. The targets ceased to govern because another statute took their place in practice and in law. A reader who describes the events of 1990 as a repeal compresses that sequence into a single act of erasure and loses the most instructive part of the sequence, which is that the correction arrived in the form of a replacement architecture with its own caps, its own scoring, and its own sequestration backstop. Supersession preserved the idea of enforcement while discarding the object that enforcement had originally claimed to govern.

Abandonment of the schedule meant that the later steps in the revised table never operated at all. The 1987 repair had set a maximum amount of $28.0 billion for fiscal 1992 and zero for fiscal 1993, with the fiscal 1993 step carrying the promise of balance that the original 1985 schedule had assigned to fiscal 1991. Neither amount functioned as a test under the successor design, because the Budget Enforcement Act framework had replaced the test before fiscal 1992 began. The revised balance date therefore passed as a date on a superseded schedule rather than as a deadline with legal force. What arrived in fiscal 1993 was not balance under any description of the record. The fiscal 1992 unified figure in the Office of Management and Budget Historical Tables was $290.3 billion, and the fiscal 1993 figure in that same series was $255.1 billion.

Did the statute ever balance the budget?

No. The fiscal 1993 balance date arrived with a unified deficit of $255.1 billion in the Office of Management and Budget Historical Tables, and fiscal 1992 closed at $290.3 billion. The fixed targets for those years never operated after the 1990 replacement of the enforcement design.

The record after 1990 belonged to the successor framework, and that framework proved durable in a way that the fixed schedule had not. The Omnibus Budget Reconciliation Act of 1993, enacted as P.L. 103-66 and signed on August 10, 1993, extended the Budget Enforcement Act design. The Budget Enforcement Act of 1997, enacted within P.L. 105-33 and signed on August 5, 1997, extended the design again. Sequestration enforcement under the successor system operated until 2002. The caps and PAYGO system survived through repeated extension across administrations and Congresses of differing composition, because compliance with a cap or an offset requirement asked officials to adjust bills rather than to deliver a macroeconomic result.

Balanced budgets emerged in the late 1990s under the successor framework combined with economic conditions. The phrasing of that record requires care in both directions. The emergence of balance under the caps and PAYGO system did not prove that caps and PAYGO alone produced balance, because the same record joined the framework to economic conditions that raised revenue and restrained some categories of spending pressure. The emergence of balance under the successor framework also did not retroactively validate the Gramm-Rudman-Hollings schedule, because balance arrived after the schedule had been abandoned and under rules that the schedule’s failure had prompted Congress to write. The sequence supports a narrower and stronger observation about institutional learning. A decade of experience with an unenforceable object taught the legislative branch to select an enforceable object, and balance followed under the enforceable object when economic conditions cooperated.

The lesson of the endgame can be stated as a namable claim, and the claim should be stated plainly before the qualifications that surround the claim receive attention. A legislature cannot legislate an outcome. The Gramm-Rudman-Hollings target was a number, and that number depended on the economy. Revenue followed employment, wages, profits, and asset values. Spending followed caseloads, interest rates, and obligations that earlier statutes had created. A recession moved the number without any member casting a vote to move the number. A forecast revision moved the estimated number that triggered the sequester without any change in enacted law at all. The operating record had demonstrated the point early, when the fiscal 1986 outcome came in more than $50 billion above the forecast path embedded in the law, and the record demonstrated the point repeatedly as actual results in fiscal 1988, fiscal 1989, and fiscal 1990 each exceeded the revised amounts for those years by margins that no procedural adjustment within the statute had closed. A procedural rule cannot bind a figure that a recession or a forecast revision moves, because the rule governs the conduct of the institution that wrote the rule, and the figure answers to forces outside that institution as well as to decisions inside that institution.

The 1990 replacement constrained legislated actions for exactly that reason, and the correspondence between the diagnosis and the prescription gives the successor statute its analytical clarity. Discretionary caps governed the appropriations that Congress wrote each year. PAYGO governed the scored effect of changes to revenue law and mandatory spending law that Congress enacted. Each instrument had a trigger that a legislator could identify before a vote, a baseline against which the effect of the vote could be measured, and a remedy that fell on the category in which the breach had occurred. None of that machinery promised a particular size for the annual shortfall, and the absence of such a promise was a feature of the design rather than an oversight. The architects of the 1990 system had watched a promised outcome fail through five fiscal years, two target schedules, one Supreme Court decision that removed the original reporting officer, and a series of evasions that included optimistic assumptions, asset sales counted as savings, payment dates shifted across a year boundary, and emergency designations. A design that promised less in terms of outcomes could demand more in terms of conduct, because the demand attached to conduct that the institution actually controlled.

The counter-reading of the statute deserves carriage in full, because the record supports the counter-reading in several particulars and because a verdict of simple futility cannot account for the legislative history that followed enactment. The statute concentrated congressional attention on the fiscal position for the first time as an organizing fact of the legislative calendar. Before Gramm-Rudman-Hollings, no mechanism forced the size of the shortfall itself onto the agenda each year in the form of a report, an estimate, and a possible automatic consequence. Budget resolutions under the Congressional Budget Act of 1974 had created a process for considering revenue and spending together, and Gramm-Rudman-Hollings overlaid that process with a numerical test that recurred on fixed dates. Committees, party leaders, and Presidents had to position themselves in relation to a number that the statute published in advance and that reporting officers measured on a schedule. Attention of that kind does not guarantee results, and the results under the schedule fell short in every year except the estimate-based clearance of fiscal 1987, but attention altered the environment in which fiscal legislation moved for the remainder of the decade.

The sequester threat also shaped enacted packages in ways that a focus on the missed targets alone can obscure. The threatened sequester after the stock market crash of October 19, 1987 framed the summit that produced a two-year package described as $76 billion of reduction, implemented through the Omnibus Budget Reconciliation Act of 1987 as P.L. 100-203 together with an omnibus appropriations act as P.L. 100-202. The fiscal 1990 sequence produced a sequester that took effect on October 16, 1989, followed by a reconciliation settlement in the Omnibus Budget Reconciliation Act of 1989, P.L. 101-239, that rescinded the order in part and preserved a portion of the automatic reductions in reduced form. The fiscal 1991 sequence produced the Andrews summit and the five-year $500 billion agreement. In each case the parties negotiated in the shadow of a threatened or actual sequester, and in each case the enacted result differed from the result that the ordinary process had been producing without that shadow. One of the authors of the act stated the intention behind the design in terms recorded by the Congressional Research Service, observing that the objective was never to trigger the sequester and that the objective was to have the threat of the sequester force compromise and action. Judged against that stated intention rather than against the statutory table, the mechanism performed a function in 1987, in 1989 and 1990, and in the terminal negotiation of 1990, even as the table itself failed in every one of those years.

The strongest element of the counter-reading concerns the relationship between the failed statute and the statute that replaced the failed one. The failure produced the successor design. The Budget Enforcement Act enforced legislated actions rather than outcomes precisely because the Gramm-Rudman-Hollings decade demonstrated that an outcome target cannot be enforced. That demonstration had a cost, paid in a constitutional adjudication, in repeated target revisions, in scoring practices adopted to satisfy estimates, and in a final estimate far beyond the trigger that the automatic remedy collapsed. The demonstration also had a product, visible in statutory text that Congress enacted in 1990 and extended in 1993 and again in 1997. A wasted failure leaves nothing behind except the record of the attempt. An instructive failure leaves a corrected design that other legislators can operate, and the caps and PAYGO framework operated through 2002 and coincided with the arrival of balance in the late 1990s. The statute therefore reads as an instructive failure rather than a wasted one, and the instruction lies in the specificity of the error. Congress learned not merely that deficit reduction was difficult, which members already understood, but that the object selected for enforcement determined whether enforcement could function at all.

The series thesis thread running through the legislation covered in this series finds in Gramm-Rudman-Hollings an unusually clear specimen. The flaw was analytically clean. Congress selected a variable that Congress did not control, attached an automatic penalty to deviations in that variable, exempted more than half of federal spending from the penalty base, and expected the penalty to compel the result that the variable refused to deliver. Each element of that construction can be isolated in the statutory text and traced through the operating record, from the estimate-based clearance of fiscal 1987 through the crash summit of that autumn, through the partial sequester and partial rescission of fiscal 1990, to the October 1990 estimate that implied reductions of 34.7 percent and 31.6 percent in the two categories. The variable that failed can be identified, the mechanism of failure can be reconstructed, and the correction can be checked against the diagnosis.

The correction stands visible in the successor statute for any reader who compares the two texts. Where the 1985 act and the 1987 repair published a table of maximum amounts descending to zero, Title XIII of P.L. 101-508 published caps on discretionary spending and a pay-as-you-go discipline for legislated changes. Where the earlier design asked whether the estimated outcome had obeyed the table, the later design asked whether enacted appropriations had obeyed a cap and whether enacted changes had paid for themselves under the scoring rules. The balance date of fiscal 1993 arrived with a unified figure of $255.1 billion in the Office of Management and Budget Historical Tables, and balance itself arrived later in the decade under the corrected framework and favorable conditions. A statute that commanded an outcome missed the outcome on the date the statute named. A statute that disciplined actions governed for more than a decade and accompanied an outcome that the earlier statute had promised and failed to deliver. The design flaw was analytically clean, and its correction remains visible in the successor statute.

The target and outcome table

The table records the statutory ceiling for each fiscal year, the recorded outcome in the unified deficit series of the Office of Management and Budget Historical Tables, and the way the sequestration question was resolved in that year. Original ceilings governed the first two years, and revised ceilings governed the years after the repair statute, with the final two revised ceilings never operating after the successor design displaced the schedule.

Fiscal year	Statutory target ($B)	Actual unified deficit ($B)	How the sequester was avoided or applied
1986	171.9	221.2	Capped sequester of $11.7B took effect March 1, 1986
1987	144.0	149.7	Estimate inside target plus $10B margin; no sequester
1988	144.0	155.2	Threatened $23B sequester replaced by post-crash summit package (OBRA 1987)
1989	136.0	152.6	Summit package second year; no sequester
1990	100.0	221.0	Sequester took effect October 16, 1989; OBRA 1989 rescinded and reduced it
1991	64.0	269.2	Threatened sequester displaced by the 1990 summit agreement and the Budget Enforcement Act
1992	28.0	290.3	Targets superseded by Budget Enforcement Act caps and PAYGO
1993	0	255.1	Targets superseded; balance date arrived under the successor design

The design lesson of Gramm-Rudman-Hollings turns on the difference between a number and an act. A maximum deficit amount described a condition of the economy as much as a decision of Congress, because revenue followed growth, employment, and asset values, and spending followed obligations that earlier statutes had created, caseloads, and interest on debt already issued. Enforcement attached to that number therefore punished or excused officials for movements that no vote had caused, and the estimate that carried the trigger could shift with a forecast revision before any legislator had acted. Caps on discretionary spending and pay as you go discipline for legislated changes reversed the attachment. Each successor instrument asked whether an identifiable bill had breached an identifiable limit, and the comparison captured in sequestration versus PAYGO explains why the successor framework proved operable across repeated extensions while the fixed schedule required revision within two years and replacement within five.

The counter-reading gives the failed schedule a durable place in budget process reform, because the statute forced the size of the shortfall onto the legislative calendar as a recurring reported fact, framed the crash summit package after October 1987, framed the partial sequester bargain for fiscal 1990, and framed the five year agreement negotiated at Andrews Air Force Base. None of those settlements attained the printed ceiling for the year in question, and the balance date arrived with a unified deficit of $255.1 billion in the Office of Management and Budget Historical Tables, yet each settlement differed from the drift that had produced a resolution requiring $55.5 billion of reduction legislation that never arrived. The failure also produced the diagnosis from which the successor design was drawn, a sequence that marks the episode as instructive rather than wasted. Readers assembling the reports, targets, court citations, and summit figures into a research file can keep your statute notes, citations, and case chronologies together free on VaultBook while tracing how a target regime gave way to a conduct regime.

Frequently Asked Questions

Q: What did Gramm-Rudman require?

The statute required the federal deficit to descend on a schedule written into law. The Balanced Budget and Emergency Deficit Control Act of 1985 set a maximum amount of $171.9 billion for fiscal 1986, followed by $144.0 billion, $108.0 billion, $72.0 billion, and $36.0 billion, reaching balance in fiscal 1991. An estimate prepared before a fiscal year began was tested against the ceiling for that year, with a $10 billion margin in the intervening years. An estimate beyond the ceiling plus the margin produced sequestration, the automatic cancellation of budgetary resources, split equally between defense and nondefense categories and applied as a uniform percentage within each category. More than half of federal spending stood exempt or protected, and war or recession allowed suspension through a joint resolution under stated growth tests.

Q: Why was Gramm-Rudman attached to a debt limit bill?

The Senate added the Gramm-Rudman-Hollings amendment to H.J.Res. 372, a measure raising the statutory debt limit to $2.079 trillion, because borrowing authority had to pass and a freestanding deficit bill could be delayed or defeated without collateral consequence. Attachment forced both chambers to confront the targets as part of a measure whose failure would leave the debt limit unchanged. Voting against the combined measure meant accepting responsibility for blocking the increase, and voting for the increase meant accepting the targets and the sequester. Conference and an amendment exchange between the chambers settled the text, the House approved the measure by 271 to 154, the Senate approved the measure by 61 to 31, and President Ronald Reagan signed the measure on December 12, 1985.

Q: What did Bowsher v. Synar decide about Gramm-Rudman?

In Bowsher v. Synar, 478 U.S. 714 (1986), the Supreme Court held by 7 to 2 that the functions assigned to the Comptroller General under the act amounted to execution of the laws. The Comptroller General evaluated the joint report of the Office of Management and Budget and the Congressional Budget Office and reported the determination on which the presidential sequestration order rested. Because the Comptroller General was removable by Congress, by joint resolution for inefficiency, neglect of duty, or malfeasance, or by impeachment, Congress could not vest execution in that officer. The automatic mechanism therefore fell. The deficit targets themselves were not struck down, and Congress retained the expedited fallback procedure and the ordinary legislative process for pursuing the schedule.

Q: Did Gramm-Rudman balance the budget?

No. The original schedule promised balance in fiscal 1991, and the repair statute of 1987 moved the balance date to fiscal 1993, with a target of zero for that year and no margin. The revised targets for fiscal 1992 and fiscal 1993 never operated, because the Budget Enforcement Act of 1990 replaced the fixed target test with discretionary spending caps and PAYGO enforcement before those years began. The balance date therefore arrived under the successor design. The fiscal 1992 unified deficit in the Office of Management and Budget Historical Tables was $290.3 billion, and the fiscal 1993 unified deficit in that series was $255.1 billion. Balanced budgets emerged later in the 1990s under the successor framework combined with economic conditions.

Q: How did Congress evade Gramm-Rudman targets?

The record described several techniques, each operating on the estimate rather than on the longer run cost of government. Optimistic economic assumptions raised projected receipts and lowered projected outlays on paper. The $10 billion margin functioned in practice as a standing upward revision of each target. Asset sales, including the sale of Conrail and sales of government loan assets, concentrated one time receipts in the target year. Payment dates shifted across the fiscal year boundary moved outlays between years. Funding shifted within appropriations, and emergency and disaster designations changed the measured total. When those techniques could not close the gap, Congress revised the targets in 1987 and abandoned the schedule in 1990, the terminal technique available to a legislature that had set the standard by statute.

Q: What replaced Gramm-Rudman?

Title XIII of the Omnibus Budget Reconciliation Act of 1990, Public Law 101-508, signed on November 5, 1990 by President George H. W. Bush, enacted the Budget Enforcement Act of 1990 and replaced the fixed deficit targets with two instruments aimed at legislated actions. Discretionary spending caps limited the appropriations that Congress enacted each year. PAYGO, or pay as you go enforcement, scored legislated changes to revenue and mandatory spending law and required offsets for changes that increased the deficit under the scoring rules. The fixed targets were abandoned and superseded rather than formally repealed. The successor design was extended by the Omnibus Budget Reconciliation Act of 1993 and by the Budget Enforcement Act of 1997, and sequestration enforcement under that design operated until 2002.

Q: Why do fixed deficit targets fail like Gramm-Rudman?

A fixed deficit target names an outcome, and an outcome depends on forces that a legislature does not control through the target itself. Revenue follows economic growth, employment, wages, profits, and asset values. Spending follows caseloads, interest rates, and obligations created by earlier law. A recession or a forecast revision can move the estimated deficit across a trigger line without any vote, and the fiscal 1986 outcome arrived more than $50 billion above the forecast path embedded in the statute. Enforcement attached to such a number also invites techniques that move the estimate, including optimistic assumptions, asset sales, and payment timing, and a legislature retains the power to revise the number when obedience becomes too costly, as the revisions of 1987 and the replacement of 1990 demonstrated.

Q: Who were the sponsors of Gramm-Rudman?

The sponsors were Senator Phil Gramm, a Republican of Texas, Senator Warren B. Rudman, a Republican of New Hampshire, and Senator Ernest F. Hollings, a Democrat of South Carolina. The sponsorship crossed party lines, with two Republicans and one Democrat lending their names to a measure that restrained a Republican administration and Congress alike, and the statute took its popular name, Gramm-Rudman-Hollings, from the three senators. One of the authors, quoted in the Congressional Research Service report on statutory budget controls, stated the purpose of the design directly, observing that the objective was never to trigger the sequester and that the objective was to have the threat of the sequester force compromise and action. Hollings later distanced himself from the record in reporting published in 2013.

Q: How did the margin change the line that triggered a sequester?

The margin was the amount by which an estimated deficit could exceed a statutory target without triggering a sequester. Fiscal 1986 and the balance year carried no margin, while the intervening years carried a margin of $10 billion, so the operative trigger levels stood at $154 billion for fiscal 1987 under the original schedule and at $74 billion for fiscal 1991 under the revised schedule. The allowance acknowledged that the test ran on a projection and that a projection could miss by a modest amount. In operation, planners aimed at the trigger level rather than at the printed ceiling, and an analysis published in National Affairs in 2013 described the margin as a standing upward revision of each target.

Q: How did the exemption list change the burden on unprotected programs?

The exemption list removed Social Security, Medicaid, veterans compensation and pensions, Aid to Families with Dependent Children, Supplemental Security Income, WIC, food stamps, the postal service fund, and the earned income tax credit from the sequestration base, with net interest exempt in operation. More than half of federal spending was exempt or protected, so a required dollar reduction had to be collected from a much smaller remainder. Special rule programs such as Medicare faced caps of 1 percent in 1986 and 2 percent in later years, contributing below the uniform percentage and leaving a further share to uncapped accounts. The percentage for unprotected programs therefore rose from two directions, and annually appropriated programs absorbed a concentration that the headline percentage understated.

Q: What happened under the capped first sequester?

Fiscal 1986 operated under a special rule for a year already under way at enactment. The sequester was capped at $11.7 billion, the order took effect on March 1, 1986, and a seven twelfths proration scaled the cancellation to the portion of the year that remained. Realized reductions reached approximately 4.3 percent for nondefense programs and 4.9 percent for defense programs, and the Education Department reduction was $170.9 million. A three judge district court had invalidated the reporting provisions on February 7, 1986 in Synar v. United States, 626 F. Supp. 1374 (D.D.C. 1986), and stayed the ruling pending Supreme Court review, so the cancellations proceeded on schedule under a judgment that had questioned the procedure authorizing the cancellations.

Q: What was the fallback procedure after the automatic mechanism fell?

Section 274(f) of the act supplied a fallback procedure designed for the judicial outcome that occurred. A Temporary Joint Committee on Deficit Reduction, made up of members of the Budget Committees, would receive the report of the Office of Management and Budget and the Congressional Budget Office. The process would then lead to a joint resolution that Congress voted on under expedited rules, compressing debate and amendment. If the resolution passed and was signed, the resolution served as the basis for a presidential sequestration order. The fallback preserved part of the forcing effect through speed, but expedition operated as internal procedure. Reductions followed only after the elected branches had acted in their assigned capacities, Congress by voting and the President by signing and ordering, which gave the procedure the constitutional safety that the automatic mechanism lacked.

Q: What changes did the reaffirmation repair make to the schedule and the trigger?

The Balanced Budget and Emergency Deficit Control Reaffirmation Act of 1987, Public Law 100-119, signed on September 29, 1987, rebuilt the trigger around the Director of the Office of Management and Budget, an executive officer, for fiscal 1988 through fiscal 1993. The Congressional Budget Office reported to the budget office and to Congress, and the budget office had to use the technical assumptions and methodologies of the Congressional Budget Office and explain any discrepancy. The balance date moved from fiscal 1991 to fiscal 1993, and the revised targets were $144.0 billion, $136.0 billion, $100.0 billion, $64.0 billion, $28.0 billion, and zero. The President received an option to exempt military personnel accounts if the same total defense reduction was achieved, subject to congressional approval, and an expedited joint resolution could supplant a final order.

Q: How did the market crash summit displace a threatened sequester?

After the stock market crash of October 19, 1987, a threatened sequester of about $23 billion framed a budget summit between the White House and Congress. The amended act capped a fiscal 1988 sequestration at $23 billion in outlay reductions, split at $11.5 billion for defense and $11.5 billion for nondefense, and the Congressional Budget Office calculated uniform percentages of 10.4 percent for defense and 8.7 percent for nondefense. A sequestration order issued in November 1987, and the summit agreement displaced the scheduled cuts with a package described as $76 billion of deficit reduction over two years, reducing the fiscal 1988 deficit by about $30.2 billion and providing nearly $46 billion of reduction for fiscal 1989. The Omnibus Budget Reconciliation Act of 1987 and an omnibus appropriations act implemented the settlement, and the November order was canceled after enactment.

Q: How was the sequester that took effect in October partly rescinded?

Lawmakers missed the October deadline for action that would have displaced an order for fiscal 1990, and a sequestration order issued on October 16, 1989 cut discretionary programs by 5.3 percent. The Education Department recorded a reduction of about $949 million on paper, with practical effects delayed because most education programs were forward funded. The Omnibus Budget Reconciliation Act of 1989, Public Law 101-239, signed on December 19, 1989, answered with an estimated $14.7 billion of deficit reduction. Title XI of that act rescinded the October order and directed a new order with adjusted and reduced reductions, prohibited any double reduction in Medicare, and legislated the treatment of transfers across the fiscal year boundary. The partial sequester largely survived in reduced form rather than ending through a full restoration.

Q: What estimate forced the endgame negotiation at Andrews Air Force Base?

The final sequestration report of the Office of Management and Budget in October 1990 estimated a fiscal 1991 deficit of $147.3 billion, against a revised target of $64.0 billion and a trigger level of $74 billion after the $10 billion margin. Compliance through the automatic formula would have required defense outlay reductions of $41.7 billion, calculated at 34.7 percent, and nondefense reductions of $41.7 billion, calculated at 31.6 percent. Reductions approaching one third of each affected base presented a consequence that neither branch treated as an operating plan. The prospect produced the budget summit at Andrews Air Force Base and a five year agreement for $500 billion of deficit reduction, implemented by the Omnibus Budget Reconciliation Act of 1990, which replaced the target regime through the Budget Enforcement Act in Title XIII.

Q: What did cancellation of budgetary resources mean in practice?

Sequestration meant the automatic cancellation of budgetary resources, meaning budget authority and related resources that law had made available for obligations and outlays. Cancellation removed legal capacity to spend rather than merely recording an intention to spend less. The required total was split equally between defense and nondefense categories, and a uniform percentage applied within each category, so each covered account lost the same proportion of resources regardless of merit, fixed obligations, or timing. Agencies then translated a percentage into operations under contracts already signed, grant commitments, procurement schedules, and personnel obligations that a percentage did not cancel. Forward funded programs added a lag, because a reduction recorded in one fiscal period reached institutions in a later funding cycle, separating the legal date of an order from the date of visible effect.

Q: Why was Social Security protected from sequestration?

Social Security stood in the fully exempt group named in the statute, so a sequestration order left the program budgetary resources untouched by cancellation. The 1985 act also placed Social Security off budget beginning in 1986, a separation summarized in Senate Report 105-3, giving the program a budgetary status apart from the totals that the enforcement machinery measured. Exemption and off budget placement together made Social Security the program least connected to the mechanism. The arithmetic effect was large, because removal of a program of that scale shrank the base across which any required reduction was divided and raised the percentage borne by unprotected accounts. The statute listed the protection without announcing a motive in the text, and the National Affairs analysis published in Spring 2013 read the exemption list as a concession that formed the price of passage in the Democratic House.

Q: What is the difference between targets being superseded and targets being repealed?

A repeal erases a statutory framework through a direct act of removal, while supersession displaces a framework by enacting a replacement that governs in practice and in law. The fixed targets met the second fate. Congress enacted the Budget Enforcement Act of 1990 as Title XIII of the Omnibus Budget Reconciliation Act of 1990, and the new architecture of discretionary caps and PAYGO took the place of the target test as the operative system of budget control. The revised targets for fiscal 1992 and fiscal 1993 ceased to govern because another statute had replaced the test, not because a repeal had struck the schedule from the books. The distinction preserves the instructive sequence of the period, in which correction arrived as a redesigned enforcement system with its own scoring and its own sequestration backstop.

Q: What did the episode teach the drafters of the successor enforcement act?

The episode taught that the object selected for enforcement determined whether enforcement could function. A decade of operation showed an outcome target moving with the economy, with forecast revisions, and with scoring techniques that included optimistic assumptions, asset sales counted as savings, payment dates shifted across a year boundary, and emergency designations. The drafters of the Budget Enforcement Act of 1990 therefore attached enforcement to legislated actions, through caps on discretionary appropriations and PAYGO discipline over enacted changes to revenue and mandatory spending law. Each instrument had a trigger identifiable before a vote, a baseline for measuring the effect of the vote, and a remedy falling in the category of the breach. Balance emerged later in the 1990s under that framework combined with economic conditions, after the target statute had supplied the demonstration that produced the corrected design.