GST legislative history is usually told as a story about delay, as if a sound reform sat in a drawer for seventeen years while politicians dithered. The passage record tells a different and more precise story. A goods and services tax levied on the destination principle does not merely simplify a tax system; it moves revenue. Under the arrangements it replaced, the state where a good was produced collected tax at the point of origin through instruments such as the central sales tax, while under a destination-based tax the revenue travels to the state where the good is consumed. Producing states therefore stood to lose, in identifiable amounts, from the first day of the reform, and consuming states stood to gain. Every year of the seventeen was spent arguing about that transfer: who would bear it, who would insure against it, and what the insurance would cost.

The scale of what was being negotiated should be fixed before the chronology begins. On the usual count, the tax that resulted subsumed some seventeen Union and state levies, among them the Union’s excise duty and service tax and the states’ value added tax, central sales tax, entry tax, octroi, purchase tax, luxury tax and entertainment tax, and it replaced them with a single levy on the supply of goods and services, collected at destination and credited through the chain. For the Union, the surrender was of two of its three great indirect taxes. For each state, the surrender was of the tax that supplied the largest share of its own revenue, set at rates its legislature had chosen, year by year, since the Constitution began. No other Indian reform has asked so many governments to give up so much of their fiscal substance at one stroke, and none had previously been attempted by a method that required each of them to agree twice, once in Parliament’s ratification arithmetic and once in their own assemblies.

That is why the legislative history of the goods and services tax is not a footnote to the tax itself but the explanation of its final shape. The tax that arrived on 1 July 2017 carries the marks of the bargain that produced it. Alcohol for human consumption sits outside it entirely. Five petroleum products sit inside its constitutional frame but outside its levy, waiting on a decision of the GST Council. A compensation promise of unusual generosity, written into the Constitution itself, carried the states through the first five years. The Council that governs the tax decides by weighted votes arranged so that neither the Union nor the states can be outvoted by the other side acting alone. None of those features was a technician’s first choice. Each was a price, negotiated at a specific stage, recorded in a specific text, and this article follows the sequence stage by stage: the expert proposal at the turn of the century, the state-level rehearsal that preceded it, the committee room in which the tax was designed before any bill existed, the first amendment bill that lapsed, the second that passed, and the concessions by which passage was bought.

The proposal: the Kelkar task forces

The idea entered the official record through committee work, not through a party manifesto. In 2002 the Union government appointed a Task Force on Direct and Indirect Taxes under the chairmanship of Vijay Kelkar, a former finance secretary and former chairman of the Tariff Commission, and charged it with recommending a thorough redesign of the tax system. The task force’s treatment of indirect taxes reached past the incremental repair of excise and sales tax and recommended a comprehensive goods and services tax: a value added tax extended across goods and services, across the Union and the states, with input tax credit flowing through the chain so that tax would fall on final consumption rather than accumulating at every stage of production and distribution. The Kelkar recommendation mattered because of its source. It was not a lobby’s demand or an academic’s model; it was the considered proposal of a committee appointed by the government of the day, and it gave every subsequent discussion a document to argue with.

A second Kelkar exercise reinforced the first. The task force appointed to advise on implementing the Fiscal Responsibility and Budget Management Act, which reported in 2004, treated indirect tax reform as part of the fiscal consolidation problem and again pointed to a goods and services tax as the structural answer to a fragmented base that leaked revenue at the borders between the Union’s taxes and the states’ taxes. The two reports together fixed the terms of debate for the following decade. The question was never again whether a destination-based value added tax was the technically preferred design; on that, the official expert record was settled early. The open questions were distributive and constitutional. A single tax on a common base required the Union and twenty-eight states to surrender separate powers they had held since the Constitution came into force, and to pool the proceeds of taxing the same transaction. No committee report could settle how that surrender would be compensated.

It is worth pausing on what the Kelkar design assumed, because the assumption became the fault line. A comprehensive GST on the destination principle is origin-neutral in theory and destination-favouring in arithmetic. A state such as Gujarat or Maharashtra or Tamil Nadu, whose factories supply the whole country, collects tax under an origin-based regime on value created within its borders and exported in finished goods. Under a destination-based regime that tax accrues where the goods are consumed, which is to say disproportionately in populous consuming states. The manufacturing states’ objection to the design was therefore not a misunderstanding of it. It was a correct reading of it, and the seventeen years that followed are unintelligible unless that is grasped at the outset. The reform’s economics were agreed early. Its incidence was the dispute.

Kelkar’s influence on the passage did not end with the task forces of 2002 and 2004. As chairman of the Thirteenth Finance Commission, which reported in December 2009, he carried the design into the constitutional machinery of fiscal transfers. The Commission recommended that the Union and the states move to what it called a flawless goods and services tax, and it put money behind the recommendation, proposing a grant of fifty thousand crore rupees to be paid to the states as an incentive and compensation pool if they adopted the model design on the Commission’s timetable. The episode matters to the passage record for two reasons. It demonstrated that compensation, as an instrument for buying state consent to tax reform, was present in the official thinking years before it became the decisive concession of 2016, and it demonstrated the instrument’s limits. The grant was offered against adoption of a model the states had not written, on conditions the states had not set, and the timetable passed without the bargain being struck. When compensation finally worked in 2016, it worked because its terms had been negotiated by the parties who would receive it, and because it stood in the Constitution rather than in a commission’s recommendation. The contrast between the two attempts is the single most instructive comparison the pre-history offers. The Finance Commission’s model also fixes, by contrast, the distance between the design as proposed and the tax as enacted. The flawless tax of the Commission’s sketch was to be levied at a single combined rate, twelve per cent on its computation, shared between the Union and the states, on a base with almost no exemptions, so that the credit chain would run unbroken from the first input to the final consumer. The tax that emerged from the negotiation has several rates, a cess, an exemption threshold, a composition scheme for small dealers, two great product exclusions and a special administrative division of the taxpayer base between the two levels of government. Every one of those departures from the model was introduced at an identifiable stage of the passage, in response to an identifiable constituency: the rates to the states’ revenue anxieties and the political economy of taxing necessities and luxuries alike, the exclusions to the treasuries of the states, the threshold and composition scheme to the small trader whose compliance capacity the Empowered Committee’s members represented. The legislative history is, in this aspect, the record of a model’s successive accommodation to the federation that had to enact it.

The system the reform had to replace

The pressure that produced the proposal came from the visible incoherence of the indirect tax system the country had assembled by accretion. At the Union level, excise duty fell on manufacture, so that the tax attached at the factory gate and embedded itself in price through every subsequent transaction without full credit. A service tax, introduced in 1994 on a handful of services, had grown service by service until it covered most of the service economy, but it ran in a separate channel from the tax on goods, so that transactions combining goods and services, from restaurant meals to works contracts, generated classification disputes as a matter of routine. At the state level, sales tax and then VAT fell on the sale of goods, entry tax fell on goods entering a local area in several states, octroi survived in Maharashtra’s municipal corporations, and a lattice of cesses, surcharges and turnover taxes filled the gaps. Between the two levels, the central sales tax, levied by the Union under a 1956 statute but collected and retained by the state from which the goods moved, taxed inter-state sales on an origin basis with no credit to the purchasing dealer, so that the tax on goods crossing a state border cascaded in full and rewarded businesses for structuring their distribution around state boundaries rather than around their customers.

The costs of this architecture were not abstract. Tax paid at earlier stages hid inside prices, so that Indian manufactures carried an embedded tax load that exports could not fully shed and domestic buyers could not see. Checkposts at state borders, where trucks queued for the verification of forms that established whether central sales tax had been paid and at what rate, functioned as internal customs posts in a country whose Constitution promised freedom of trade and commerce throughout its territory. Businesses organised warehouses state by state to convert inter-state sales into stock transfers and local sales, a distortion of logistics so thorough that the warehouse map of Indian industry was, in effect, drawn by the tax system. Successive official committees, beginning with the tax reforms committee chaired by Raja Chelliah in the early nineteen nineties, had identified the cure as a value added tax extended across goods and services and across both levels of government. The Kelkar reports restated that cure in its complete form. What none of the committees could supply was the consent of the governments whose revenues the cure rearranged, and supplying that consent is the whole of the legislative history that follows.

The rehearsal: states learn value added tax

A national goods and services tax could not have been legislated in 2002 even if the politics had allowed it, because the administrative precondition did not exist. State taxation of goods still ran through first-point and multi-point sales taxes with no general credit for tax paid on inputs, so the value added mechanism that GST generalises had never been operated by the state administrations that would have to run it. The rehearsal came through the state-level value added tax, and the institution that conducted it was the second protagonist of this history.

In November 1999 a conference of chief ministers and state finance ministers, convened by the Union finance minister, set up a standing committee of state finance ministers to deliberate on the design of a state-level VAT. That body, which became the Empowered Committee of State Finance Ministers, was not a Union department and not a parliamentary committee. It was a forum of the states themselves, serviced by a secretariat and chaired by one of their own number, and it acquired through the VAT exercise a function no Indian institution had performed before: it negotiated a common tax design among sovereign-level governments that retained the legal power to refuse it. Its first chairman, Asim Kumar Dasgupta, finance minister of West Bengal, held the office from the Committee’s formation until July 2011, a tenure that spanned the entire VAT transition and the first decade of GST design, and under him the Committee produced the White Paper on state-level VAT in January 2005 that fixed the common structure states would adopt.

Haryana moved first, introducing VAT in 2003. The large majority of states followed together on 1 April 2005, replacing their sales taxes with a multi-point tax on value added, with input tax credit, floor rates agreed through the Empowered Committee, and a common framework of registration, returns and assessment. Uttar Pradesh, the last holdout, adopted VAT only from 1 January 2008. The White Paper’s structure is worth a line of detail, because it was the template for everything the Committee later attempted with GST. It proposed a small number of common rates across all states, principally a lower rate of four per cent and a standard rate of twelve and a half per cent, with a short list of exempt goods and a higher rate for a narrow class of demerit goods, and it bound the participating states to a floor-rate discipline under which no state would undercut the agreed rates to poach trade from its neighbours. The discipline was not perfect; states deviated at the margins, added local levies, and fought over the treatment of particular commodities. But the episode established that a common rate structure could be negotiated without any legal power to compel it, held together by the mutual interest of the participants and by the publicity of defection. When the same ministers later faced the GST design, they were not theorising about cooperative federalism. They had operated it, imperfections included, for half a decade. The transition was not administratively smooth, and its political costs are part of the inheritance. Traders’ organisations resisted the new tax in several states with strikes and shutdowns in the adoption years, on complaints that ranged from the burden of computerised invoicing to the fear that input tax credit would expose the whole of a dealer’s purchases to the record. State governments bore those costs at the retail end of politics, in their own constituencies, for a reform whose design had been negotiated nationally. The experience cut both ways in the later GST negotiation. It proved to the states that a value added tax could be administered and that revenues, after the transition, grew; the buoyancy of VAT collections in the years after 2005 was the single most persuasive exhibit in the Empowered Committee’s later advocacy of GST. It also taught every finance minister who lived through it that the government which introduces a tax reform owns its transitional pain in the public mind, however collective the design. The caution with which the states approached the far larger surrender of GST, and their insistence that the transitional risk be underwritten by the Union in advance and in writing, is partly the memory of 2005 speaking.

The transition taught three lessons that GST would inherit. The first was mechanical: dealers, accountants and state tax administrations learned to operate an invoice-credit system, and the information trails that system created turned out to raise collections in ways the old sales tax never had. The second was institutional: states discovered, through the Empowered Committee, that they could negotiate a shared design, police each other’s deviations through peer pressure, and present the Union with something close to a collective state position. The third was fiscal and cautionary: the states that moved to VAT had been promised compensation for transitional revenue loss, the Centre had paid it on a tapering schedule, and arguments about the adequacy and duration of that compensation left a deposit of distrust that every later GST negotiation would draw on. When the states were later asked to surrender far more than sales tax, they negotiated as parties who had already once been promised that a reform would not cost them money, and who intended this time to have the promise in writing.

The design years: the Empowered Committee and the First Discussion Paper

While VAT bedded down, the same committee room turned to the larger design. In the Union Budget for 2006-07 the finance minister, P. Chidambaram, announced that India would move to a goods and services tax with effect from 1 April 2010, and the Union Budget for 2007-08 asked the Empowered Committee to work with the Union government to prepare the road map. The Committee responded in May 2007 by constituting a Joint Working Group of Union and state officials, co-convened by the adviser to the Union finance minister and the Committee’s member-secretary, which divided itself into sub-groups, consulted industry and experts, and reported on 19 November 2007. Its central recommendation fixed the architecture that survives in the enacted tax: a dual GST, levied concurrently by the Union and by each state on the same base, with an integrated tax on inter-state supplies collected by the Union and apportioned, rather than a single national tax administered from Delhi. The dual structure was itself the first great concession, made before any bill was drafted. A unitary GST would have been simpler to administer and cleaner in design, and it was never seriously available, because it would have required the states to hand their largest revenue source to the Union outright. Dual GST was the federal price of the concept, and it was paid at the design stage, in committee, years before Parliament saw a text.

When did GST move from an idea to a dated commitment?

The Union Budget for 2006-07 announced that a goods and services tax would be introduced with effect from 1 April 2010, and the following year’s Budget commissioned the Empowered Committee to prepare the road map. From that announcement the reform had a public deadline and a named designing institution, rather than remaining an expert recommendation awaiting adoption.

The design work culminated, for that phase, in the Empowered Committee’s First Discussion Paper on the goods and services tax, released in New Delhi on 10 November 2009 under Dasgupta’s chairmanship. The paper set out the proposed structure in public for the first time: the dual levy, the taxes to be subsumed at each level, the treatment of inter-state trade, the threshold and composition questions, and the constitutional amendments that would be needed. It was expressly a discussion document, issued to invite the views of industry, trade, agriculture and the public, and its release by a committee of state finance ministers rather than by the Union finance ministry was the point. The states were presenting themselves, through their committee, as co-authors of the reform rather than as its subjects. The paper also exposed, by the act of publication, how much remained unresolved. Rates were not fixed. The treatment of petroleum products, alcohol and electricity was contested. The compensation question for states that lost revenue was acknowledged rather than answered. And the deadline of 1 April 2010, by then months away, was visibly unattainable, because the constitutional amendment on which everything depended had not even been introduced.

The First Discussion Paper is the most under-read document in the passage record, and it repays attention because nearly every element of the eventual settlement can be seen in it in embryo. Its model was dual and concurrent from the first page: a central goods and services tax and a state goods and services tax levied on the same transaction, with the integrated tax on inter-state supplies, which already implies the constitutional surgery the bills later performed. It proposed the subsumption of the Union’s excise and service tax and of the states’ VAT, central sales tax, entry tax, octroi and allied levies into the two components, which already identifies the municipal and local revenues that would have to be dealt with. It treated exports as zero-rated and imports as taxable, which already fixes the destination principle in the design rather than leaving it to inference. And it left open, expressly and by name, the questions the next seven years would be spent answering: the rates, the thresholds, the fate of petroleum and alcohol, and the compensation of losing states. A reader who compares the discussion paper of 2009 with the amendment of 2016 sees the whole history in the difference between the two documents. The architecture did not change. Everything that changed was distributive, and everything distributive in the final text can be traced to a question the discussion paper had the honesty to leave blank.

The Empowered Committee as an institution

The Committee deserves a closer portrait, because it is the institution on which the whole passage turned and the one least visible in the constitutional text, in which it does not appear at all. It had no statutory basis. It was created by a political decision at a conference of governments, it was serviced by a small secretariat in Delhi, its chairman was elected by its members from among the state finance ministers, and its conclusions bound nobody. Its authority was of a different kind. Because every state government sat in it, a conclusion the Committee announced was a conclusion the states had already reached among themselves, and the Union finance ministry, which attended and serviced the process, could not dismiss a Committee position as the demand of a single dissatisfied state. The Committee also worked by consensus in practice, which meant its positions tended to be the states’ minimum common terms: the point beyond which at least one government would not move. Union negotiators complained across the years that the Committee’s consensus method made it slow and its positions rigid. The complaint identifies the institution’s function. A ratifying coalition announcing its minimum terms in advance is not a defect in a negotiation that requires ratification; it is the negotiation conducted in the open.

Two features of the Committee’s life explain its longevity. The first is that it survived every change of government among its members, because its membership turned over with state elections while its secretariat, its agenda and its accumulated drafts persisted; the design work of 2007 was still the working text of 2013 because the institution, unlike any ministry, never lost its file. The second is that it gave state finance ministers a collective identity distinct from their party identities. Within the Committee, ministers of states ruled by parties in opposition at the Union habitually carried the states’ case to a Union government of their own party’s rivals, and ministers of the Union ruling party’s states resisted that same government on the states’ brief; Sushil Kumar Modi, a BJP deputy chief minister, spent two years as the Committee’s chairman pressing a Congress-led Union government for better terms, and later said that a very large share of the design consensus, some eighty per cent on his estimate, had been built in that phase. The GST Council inherited the Committee’s membership and its method, and added only the two things the Committee had lacked: a place in the Constitution and votes with weights attached.

The design questions beneath the politics

Beneath the argument between governments ran a set of technical questions that the Empowered Committee’s machinery worked on continuously, and the passage record is incomplete without them, because several of the political sticking points were technical questions wearing federal dress. Three committees of officers constituted by the Empowered Committee at Bhubaneswar in January 2013 map the terrain. One was charged with the place of supply rules and the revenue neutral rate: the rules that determine, for a transaction spanning states, which state is the destination entitled to the tax, and the rate at which the new tax would yield what the old taxes yielded. A second was charged with dual control, thresholds and exemptions: the question of which government’s officer would assess a given dealer, the turnover floor below which dealers would be left out, and the list of goods and services neither level would tax. A third was charged with the integrated tax and the tax on imports: the mechanism by which tax on an inter-state supply would be collected by the Union and settled to the destination state without breaking the credit chain.

Each question had a distributive edge. The revenue neutral rate was not one number but twenty-nine: a rate neutral for a manufacturing state with a large origin-based take was not neutral for a consuming state, and the difference between the two was, in substance, the size of the compensation problem. Dual control decided whether a small dealer would face one tax administration or two, which was a question about the compliance burden the reform would impose on the smallest businesses and about how much of their establishment each state administration would surrender to a shared roster. Place of supply decided the tax entitlement in the fastest-growing part of the economy, services consumed across state lines, where no physical crossing marked the destination. The committees of officers, Union and state officials working in the Empowered Committee’s name, narrowed these questions to options; the finance ministers chose among the options in the committee room; and the choices hardened into the model law. This division of labour, officials narrowing and ministers deciding, was the Empowered Committee’s real invention, and the GST Council inherited it entire, with its own committees of officers performing the same function for the Council after 2016.

The central sales tax wound

One element of the old system deserves separate treatment, because its phase-out rehearsed, in miniature and in bad temper, every argument the GST negotiation would later repeat. The central sales tax was an origin-based levy on inter-state sales. Because it was not creditable, its abolition was a precondition of any destination-based design: an inter-state transaction taxed at origin without credit would break the input tax credit chain at every border. All parties agreed in principle that the tax would go when GST came, and, in anticipation, the Union and the states agreed to reduce it in stages. The rate came down from four per cent to three per cent in 2007 and to two per cent in 2008, and the states were to be compensated for the revenue forgone on a declining schedule, on the shared assumption that GST would arrive on 1 April 2010 and end the matter.

GST did not arrive on that date, or on the dates that succeeded it, and the compensation question outlived the schedule designed for it. The states pressed claims for the losses of the phase-out years; the Union disputed the computation and the duration of its liability; payments lagged behind claims; and a technical disagreement about a two per cent levy hardened into a settled state conviction that the Union’s compensation promises were written to expire before the losses did. The argument mattered far beyond its amounts. When, in the GST negotiations, the Union offered compensation for the transition to the new tax, the state finance ministers evaluating the offer had a recent, concrete precedent for how such offers behaved in practice, and they priced the new offer accordingly. The demand that eventually carried in 2016, for compensation guaranteed in the constitutional text, for a defined five years, on a base year and a growth rate fixed in advance, is the central sales tax dispute translated into drafting. Each of its three features answers a specific complaint from the earlier quarrel: the text answers the expiry of assurances, the five-year term answers the argument about duration, and the pre-agreed base and growth rate answer the argument about computation. Legislative history rarely shows its causation so cleanly. The guarantee in the amendment is the scar tissue of the levy that preceded it.

The passing of the 1 April 2010 deadline, unremarked in any formal instrument, is the first entry in the reform’s long calendar of missed dates, and it set the pattern for the rest. A target announced in a budget speech binds nobody; it is a statement of intent by one government, and the reform required the concurrence of thirty. When the date passed, it was succeeded by revised expectations voiced from budget to budget, each received by the states as evidence that the Union’s urgency was rhetorical while their own exposure was real. The lesson the state finance ministers drew, and acted on in 2011 and again in 2016, was that dates in this reform would be set by the completion of the bargain and not by announcement, and that a government which named a date before it had bought the consent was merely scheduling its own embarrassment. The government of 2014 named April 2016 in the same spirit and missed it in the same way, and hit July 2017 only because by then the consent had been bought in full. The calendar of the goods and services tax is therefore not a record of sloth. It is a record of the sequence in which the federation’s decisions can actually be taken, announced dates first and real dates after the bargaining, in every phase without exception.

GST legislative history turns on a constitutional amendment

The obstacle was the Seventh Schedule. Under the Constitution as it stood, the power to tax goods and the power to tax services were divided between the Union and the states in a way that made a common tax on a common base legally impossible by ordinary legislation. The Union could tax manufacture through excise and services through its own levy; the states could tax the sale of goods within their borders; neither could tax the whole supply chain, and each legislature’s list of tax entries stopped where the other’s began. A goods and services tax, which taxes a single supply once and credits the tax paid at earlier stages regardless of which government collected it, required both levels to legislate on the same base at the same time. Only a constitutional amendment could confer that concurrent power, insert the machinery for the inter-state levy, and create the joint institution that would recommend rates and rules.

Why could GST not be created by an ordinary Act of Parliament?

The tax entries of the Union List and the State List divided the field: the Union taxed manufacture and services, the states taxed sales, and neither could levy a comprehensive tax on the supply of goods and services. A value added tax spanning the chain needed concurrent power over one base, which only an amendment under Article 368 could create.

The amendment route carried a second consequence that shaped the entire negotiation. Because the amendment altered the distribution of legislative powers in the Seventh Schedule, it fell within the proviso to Article 368 and required, after passage by both Houses of Parliament with special majorities, ratification by the legislatures of not less than half the states before it could be presented for presidential assent. The working of that requirement is set out in the article on the constitutional amendment process in India, and its significance for GST was practical rather than formal. The states were not consultees in this reform. They were, collectively, a ratifying chamber. A design that a majority of state legislatures would not ratify was not a design at all, and every state finance minister in the Empowered Committee negotiated in the knowledge that his or her assembly’s vote would eventually be needed. That is the structural reason the concessions catalogued later in this article were not optional generosity on the Union’s part. They were the purchase price of a ratification that could not be compelled.

The first attempt: the Constitution (115th Amendment) Bill, 2011

The bill, when it came, had already been fought over in draft for more than a year. Through 2010 the Union finance ministry circulated successive draft amendment texts to the states, and the drafts became the terrain on which the institutional argument was first joined. The point at issue in the draft war was not the tax but the Council: how it would be created, who would chair it, how its votes would be weighted, and whether its recommendations would bind. Drafts that located the power to constitute the Council in Parliament, or that appeared to give the Union a decisive voice in a body pronouncing on state revenues, drew formal rejection from the BJP-ruled states. Gujarat’s government stated in August 2010 that the amendment had come at a time when consensus was lacking even within the Empowered Committee, and that it fundamentally altered the powers of the states to levy and collect indirect taxes. The sequence is worth recording exactly, because it inverts the usual picture of a bill drafted by the Union and then opposed. Here the opposition preceded the bill, was conducted against texts the states had been sent for the purpose of agreement, and succeeded in changing the drafting: by the third draft, and in the bill as introduced, the Council was to be created by presidential order on the amendment’s commencement, chaired by the Union finance minister, with the states represented by their own ministers. Even that settlement did not hold the objectors, because the weighting of votes and the dispute authority remained.

The first text arrived on 22 March 2011, when the Union finance minister, Pranab Mukherjee, introduced the Constitution (One Hundred and Fifteenth Amendment) Bill, 2011 in the Lok Sabha. The Speaker referred it on 29 March 2011 to the Standing Committee on Finance. The bill followed the architecture the Empowered Committee had designed. It proposed a new article conferring concurrent power on Parliament and the state legislatures to tax goods and services, a levy on inter-state supplies collected by the Union, a definition of the goods and services tax that excluded alcoholic liquor for human consumption, and two new institutions. The first was a Goods and Services Tax Council, chaired by the Union finance minister with the state finance ministers as members, charged with recommending rates, thresholds, exemptions and the goods and services to be brought within the tax. The second was a Goods and Services Tax Dispute Settlement Authority, a body empowered to adjudicate disputes between the Union and the states, or among states, arising from deviations from the Council’s recommendations that affected revenue.

The bill’s drafting reached well beyond the new taxing articles. To make a common base workable it proposed corresponding changes across the constitutional text: to the articles governing the distribution of tax revenues between the Union and the states, to the restrictions on state taxation of inter-state sales carried in Article 286, to the definitions in Article 366, to the Sixth Schedule’s provisions for the autonomous councils of the north-east, whose fiscal arrangements could not simply be overridden, and to the Seventh Schedule entries themselves, where the Union’s excise entry and the states’ sales tax entry would be recast around the new concurrent power. It also contemplated the disappearance of the category of declared goods, the goods of special importance whose interstate taxation the Constitution had restricted since the nineteen fifties, because a destination-based credit system has no room for a separate regime of restricted goods. The breadth of the surgery explains the breadth of the scrutiny the bill received. The Standing Committee’s report ran to chapters on each of these heads in turn, and each head had a constituency: the north-eastern states for the Sixth Schedule, the commodity boards and producing states for declared goods, the municipal bodies for the local levies, entry tax and octroi among them, that the new tax would subsume without replacing their revenue line for line.

One feature of the bargain absent from the objections deserves emphasis, because it structured the states’ eventual consent. Before the amendment, the states could not tax services at all; the service tax belonged to the Union alone, levied under its residuary power, and services were the fastest-growing component of the economy and of the tax base. The concurrent power the bill proposed would give every state legislature, for the first time, a share in the taxation of services within its borders. For states whose economies were weighted toward services, and for poorer states whose goods production was thin but whose consumption of services grew with their incomes, this was a substantial acquisition, and it was understood as such in the committee room. The states were not being asked only to surrender. They were being offered entry into a tax base from which the Constitution had excluded them since 1950, in exchange for pooling the base they already held. Accounts of the negotiation that present the states purely as losers omit the half of the trade that made the trade possible.

The states’ objections formed quickly and on identifiable lines, and they were objections to the institutional design rather than to the destination principle in the abstract. The proposed Council’s voting arrangements were read in the state capitals as giving the Union a dominant voice in a body whose recommendations would in practice bind state budgets. The Dispute Settlement Authority was read as something stronger: a body, created under Union sponsorship, that could sit in judgment on a state’s decision to depart from a recommended rate and entertain complaints that the departure cost another government revenue. To state finance ministers who had spent a decade building the Empowered Committee precisely as a forum of equals, an authority of that kind inverted the relationship. The negotiation was supposed to happen among governments; the bill appeared to place a tribunal above it. Objections on that theme came most sharply from states ruled by the Bharatiya Janata Party, then in opposition at the Union level. In February 2011, weeks before the bill was introduced, the finance minister of Madhya Pradesh, Raghavji, said of a Union draft that it would destroy the states’ fiscal autonomy, and Gujarat’s finance minister, Saurabh Patel, called the draft retrograde and against fiscal federalism. Gujarat’s government, under Chief Minister Narendra Modi, pressed further objections of substance: that a destination-based tax would inflict severe losses on a manufacturing state, that the abolition of the central sales tax without a settled compensation mechanism repeated an old grievance, and that the information technology backbone for a national invoice-matching system did not exist and had to precede, not follow, the constitutional change.

Other states pressed different losses. Tamil Nadu, another manufacturing state with a powerful chief minister in J. Jayalalithaa, objected to the permanent surrender of fiscal autonomy and to the revenue consequences for producing states. West Bengal’s position shifted with its government. Consuming states of the north and east, poorer and more populous, stood to gain under the destination principle and were correspondingly readier to proceed, provided transitional losses were covered. What united the objectors, across party lines, was a demand that the compensation question be settled in the text rather than in speeches, and a refusal to accept institutional machinery in which the Union could outvote or overrule them. The bill also carried exclusions that limited the reform from the start. Petroleum products were kept outside the proposed tax, along with alcohol; the states taxed both heavily, petroleum through their sales tax and VAT and alcohol through state excise, and neither the Union nor the states were willing to pool the two richest handles of indirect taxation in the first version of a system neither had operated.

The Standing Committee on Finance, chaired by Yashwant Sinha, a former Union finance minister sitting on the opposition benches, examined the bill for over two years, took evidence from the finance ministry, from industry and from experts, and presented its seventy-third report to the Lok Sabha on 7 August 2013. Among the witnesses who appeared before it was Asim Kumar Dasgupta himself, by then out of office in West Bengal, appearing as a former chairman of the Empowered Committee to give the committee the states’ institutional memory of the design. The detail captures something real about the passage: the scrutiny of the first bill was conducted by the Parliament of a government that could not pass it, chaired by a finance minister of the opposition that would later pass its successor, and assisted by the designer of the structure it was examining. The report’s authority in the later drafting derived from exactly that composition. When the drafters of 2014 treated the report as settled groundwork rather than as the previous government’s homework, they were recognising that the committee stage had already performed the examination the new bill would otherwise have had to undergo from the beginning. The report supported the reform and recommended that the bill be taken forward, but it recorded the federal anxieties in detail, recommended that the concerns of states on fiscal autonomy and compensation be addressed, and questioned elements of the design, including the dispute settlement mechanism. By the time the report arrived, the political ground had moved under the bill. The Empowered Committee, which lost Dasgupta in July 2011 on his defeat in the West Bengal assembly elections, elected Sushil Kumar Modi, Bihar’s finance minister and a BJP leader, as its chairman on 18 July 2011, and under him the states pursued their own revision of the text. A Committee on GST Design, set up after a meeting between the Union finance minister and the Empowered Committee in November 2012, reported in January 2013, and the Empowered Committee at Bhubaneswar in January 2013 and at Shillong in November 2013 successively recommended changes to the pending bill: on the Council’s composition and voting, on compensation, on the treatment of entry tax and other levies, and on the dispute authority. The Union government revised its draft more than once in response, accepting some recommendations and resisting others, and a further revised draft was still circulating between the Union finance ministry and the Empowered Committee in March 2014.

The lapse of 2014

The bill died of the calendar. The fifteenth Lok Sabha was dissolved in May 2014 on the expiry of its term and the completion of the general election, and under the rules governing pending legislation a bill introduced in the Lok Sabha and not passed by it lapses on dissolution. The Constitution (115th Amendment) Bill had never been taken up for consideration and passing. It had spent three years in committee and in correspondence between the Union and the Empowered Committee, and it lapsed with the House that had received it. The lapse is sometimes narrated as a failure of will in the government of the day, and the government of Manmohan Singh, weakened in its final years, plainly did not force the bill to a vote it might have lost. The fuller reading is that the bill could not have been carried in the form in which agreement existed, because agreement did not exist. The ratification requirement meant the bill also needed the state legislatures, and the legislatures of the objecting states were controlled by parties whose finance ministers were on record against the institutional design. A vote forced in Delhi in 2013 would have produced either a defeat in Parliament or a ratification campaign against hostile state governments. The lapse closed the first attempt, and it clarified the terms on which a second attempt would have to be made: the institutional objections would have to be answered in the text, the compensation promise would have to be written into the Constitution rather than offered across a table, and the drafting would have to be done with the states rather than presented to them. A lapse, it should be noted for precision, is not a rejection. No House voted the bill down, no clause was negatived, and the work product of the first attempt, the standing committee’s examination, the Empowered Committee’s recommended revisions, the finance ministry’s successive drafts and the Committee on GST Design’s report, survived the dissolution intact and passed into the hands of the incoming government as a nearly complete file. Constitutional amendment bills lapse frequently in Indian practice and are reintroduced when their season returns; what distinguishes this lapse is that the bill’s opponents of 2011 inherited it in 2014 as its sponsors, and treated the file not as the previous government’s failure but as their own preparation. The continuity of the text across the lapse is the strongest single evidence that the reform’s difficulties were never located in its drafting. They were located in its distribution, and distribution could only be settled by the parties who bore it, at a moment when the party holding the Union needed the amendment more than the parties holding the states feared it.

The reversal of positions

No feature of the GST passage record is more often remarked, and none better deserves the neutral statement the record supports, than the exchange of positions between the two national parties. Through the years of the United Progressive Alliance government, the Congress party held the Union finance ministry and championed the reform: Chidambaram set the 2010 target, Mukherjee introduced the 2011 bill, and Congress-led state governments broadly supported the Union drafts. In the same years the BJP, governing Gujarat, Madhya Pradesh, Chhattisgarh and other states, supplied the most organised resistance, on the stated grounds of fiscal autonomy, manufacturing-state losses, uncompensated central sales tax abolition, and unready information technology. Between May 2014 and August 2016 the occupants of the two positions changed places. The BJP, in government at the Union under Prime Minister Narendra Modi, the same leader who as Gujarat’s chief minister had resisted the amendment, introduced the successor bill within seven months of taking office and staked its economic programme on passing it. The Congress, in opposition and holding the balance in the Rajya Sabha, resisted the bill through 2015 and into 2016, demanding that a ceiling of 18 per cent on the tax rate be written into the Constitution, that the additional one per cent levy for manufacturing states be dropped, that petroleum products and alcohol be brought within the tax, and that an independent dispute settlement mechanism be restored.

The exception that proves the reading is Tamil Nadu. The AIADMK government of J. Jayalalithaa opposed the reform under the UPA on the ground that a manufacturing state would lose revenue and fiscal autonomy, and it opposed the successor bill under the NDA on the same ground, its members walking out of the Rajya Sabha rather than voting for the amended text in August 2016. No reversal occurred, because the party’s institutional position did not change: it held a producing state’s government throughout, and its argument followed its seat with complete consistency across two Union governments of opposite parties. West Bengal under the Trinamool Congress supplies the complementary case from the consuming side: initially hostile in the first years after 2011, it moved to support as the compensation terms hardened, and its finance minister, Amit Mitra, chaired the Empowered Committee through the final design phase before leading his state to ratify in the second wave. The pattern across the states is therefore not a pattern of ideology at all. Producing states with governments of every party resisted longest; consuming states with governments of every party settled earliest; and the two national parties argued whichever brief their governments’ seats assigned them.

Both reversals were denounced by the other side as opportunism, and both are better explained in federal terms. A party’s view of a fiscal bargain between the Union and the states depends on which side of the table its governments sit. The BJP’s state governments in 2011 were defending the revenues and discretion of producing states against a Union design they did not control; the BJP’s Union government in 2015 was seeking a national market and needed the states’ consent, and it obtained that consent by conceding much of what its own state governments had demanded. The Congress in 2011 was a Union government asking states to trust its assurances; the Congress in 2015 was an opposition party, strong in the upper House and with state governments of its own to protect, demanding that the assurances be converted into constitutional text and that the rate be capped against future governments. The arguments on each side did not change. Fiscal autonomy, compensation, the treatment of the richest tax handles, the fear of an uncapped rate: the same brief was argued by different advocates as the chairs were rearranged. The passage record is the clearest modern demonstration that in Indian fiscal federalism the durable interests are institutional, attaching to the level of government a party happens to hold, and not ideological in the sense the party manifestos describe.

The second attempt: the Constitution (122nd Amendment) Bill, 2014

The new Union government sent a fresh draft of the amendment to the Empowered Committee in June 2014, within weeks of taking office, and the drafting method announced the lesson learned from the lapse. The bill was to be built on the accumulated work: the Empowered Committee’s recommendations, the Standing Committee’s report on the lapsed bill, and the revisions negotiated through 2013 and into 2014. The Union Cabinet approved the proposal on 17 December 2014, and the finance minister, Arun Jaitley, introduced the Constitution (One Hundred and Twenty-second Amendment) Bill, 2014 in the Lok Sabha on 19 December 2014. The bill had twenty-one clauses, and its differences from the lapsed text mark exactly where the negotiation had moved.

Three changes defined the new bill. The Dispute Settlement Authority, the institution the states had read as a tribunal set over them, was gone; disputes were left, in the first instance, to the Council itself. Petroleum products were brought within the constitutional definition of the tax, but with their actual levy deferred: the Council would recommend the date from which GST would apply to petroleum crude, high speed diesel, motor spirit, natural gas and aviation turbine fuel, so the states’ petroleum revenues were protected for a transitional period whose length the Council itself would control. And the manufacturing states received a tangible bridge. Clause 18 of the bill provided for an additional tax on the supply of goods in the course of inter-state trade or commerce, not exceeding one per cent, to be levied and collected by the Union for two years or such longer period as the Council might recommend, and assigned to the state from which the supply originated. The additional levy was, in design terms, a defect: an origin-based charge layered on a destination-based tax, reintroducing a measure of the cascading the reform existed to remove. In negotiation terms it was a recognisable instrument, paying the producing states for the transition out of the revenues of the transition itself. Its pedigree was longer than the bill. Producing states had canvassed an origin-based supplement through the Empowered Committee years, in versions running as high as two per cent, and the finance minister defended the one per cent version in the House as a transitional comfort for manufacturing states, limited to two years in the first instance. The levy was thus the producing states’ second line of defence, behind compensation and ahead of outright resistance, and its career through the bill tracks the negotiation exactly: inserted to hold the manufacturing states in 2014, endorsed with safeguards by the select committee in 2015, and traded away in August 2016 when the compensation guarantee was made strong enough to hold them without it. The bill also carried forward the exclusion of alcohol for human consumption from the tax altogether, an exclusion no participant in the negotiation seriously contested, and it repeated the promise of compensation to the states for revenue loss, a promise the states had learned to want in harder form.

The resistance of 2011 had not wholly dissolved in the states, either. Gujarat’s government, by then led by Anandiben Patel, filed a written submission against the bill with the select committee, arguing that it attacked the states’ fiscal autonomy, that it failed to address the revenue concerns of manufacturing states, and that the abolition of the central sales tax with the move to the destination principle would inflict severe losses on a producing state. The submission is the last formal statement of the case Gujarat had made since 2010, and its fate measures the distance the negotiation had travelled: the select committee heard it, the compensation guarantee in the bill answered its central demand in kind if not yet in degree, and Gujarat ratified the amendment within weeks of its passage.

The Lok Sabha passed the bill on 6 May 2015. The Congress, whose members had pressed their amendments without success, walked out of the House before the vote, a rehearsal in the lower House of the position it would hold for the following year in the upper. In the Rajya Sabha, where the government lacked a majority, the bill was referred on 14 May 2015 to a select committee of twenty-one members chaired by Bhupender Yadav, whose examination of the bill’s committee stage is treated in the article on parliamentary committees and the scrutiny of bills. The select committee consulted state governments, experts and stakeholders and reported on 22 July 2015, and the Union Cabinet accepted the greater part of its recommendations within the week. The committee had broadly endorsed the bill, including, with safeguards, the additional one per cent levy for an initial period, and it recommended a firmer compensation arrangement, proposing that compensation be paid from a dedicated fund. Members of the Congress filed their disagreement with the committee’s direction on the points their party was pressing in the House: a rate ceiling in the constitutional text, deletion of the additional levy, and the inclusion of petroleum and alcohol within the tax from the outset. The committee stage therefore closed with the shape of the eventual settlement already visible in outline. Everything except the Congress party’s core demands had been conceded, recommended or drafted; the remaining argument was about whether the final concessions would be made, and the Rajya Sabha was where they would have to be made.

The missed April of 2016

The new government had set itself a public target of introducing the tax on 1 April 2016, and through 2015 the executive machinery worked to that date even as the constitutional bill waited on the Rajya Sabha. A goods and services tax network, the Goods and Services Tax Network, had been incorporated in 2013 as a non-profit company owned jointly by the Union, the states and financial institutions, charged with building the common information technology portal on which registration, returns and invoice matching would run for every taxpayer in the country; its chairman, Navin Kumar, a former chief secretary of Bihar, fronted a build that was itself an answer to Gujarat’s old objection that the technology had to precede the tax. The Empowered Committee, in what proved to be its final phase of design work under its later chairman, West Bengal’s finance minister Amit Mitra, closed out much of the remaining design agenda: the threshold below which dealers would stay outside the tax, the division of administrative control between Union and state officers over the same taxpayer, the treatment of the area-based exemptions, and the drafting of the model state law that every legislature would eventually enact. By the time the select committee had reported and the Cabinet had accepted its main recommendations, the design questions that had occupied the years from 2007 to 2013 were substantially settled. What remained open was narrower and harder: the Congress party’s demands in the House where it held the balance, and the manufacturing states’ additional levy, which the select committee had endorsed and the Congress demanded be deleted.

The target date passed without the bill. The winter session of 2015, in which the government had hoped to secure passage in time for an April launch, was consumed by disruption, and the budget session of 2016 went the same way on the bill. The missed April is a useful corrective to the two lazy accounts of the delay. It was not caused by unfinished design; the design was ready, the technology company was building, and the Empowered Committee had done its work. Nor was it caused by state resistance of the 2011 kind; the states, through their committee, were pressing for the tax. It was caused by the distribution of seats in one House of Parliament interacting with a credit contest between two parties, each of which had held both positions on the reform within the decade. The eventual settlement was negotiated in the weeks before the monsoon session of 2016, in direct exchanges between the finance minister and the Congress leadership, and its terms, the deletion of the levy, compensation in the text, the Council’s dispute mechanism, were the terms on which the Congress’s support, or at least the support of enough members beyond the governing coalition to reach the special majority, could be had.

The stall in the Rajya Sabha and the settlement of August 2016

Through the rest of 2015 and the first half of 2016 the bill waited in the Rajya Sabha while the arithmetic of that House governed the reform’s fate. The monsoon session of 2015 ended with the bill listed but not carried; the winter session of 2015, on which the April 2016 target depended, was lost to disruption; and the budget session of 2016 repeated the pattern on this bill while passing other legislation. In parallel, the government worked the arithmetic rather than the argument. Regional parties with state governments of their own, the Trinamool Congress, the Biju Janata Dal, the Janata Dal (United), the Samajwadi Party and others, were sounded on their terms, and since their governments’ interests ran with the states’ side of the bargain, the compensation and Council terms already in the bill answered most of them. By the middle of 2016 the Congress stood close to isolated in its resistance, pressing demands that other opposition parties would not co-sign, and its leverage had narrowed to the special majority threshold itself. The final negotiation therefore took the form it did: a bilateral settlement between the government and the Congress in the last days of July 2016, traded against the text of the official amendments, in which the Congress obtained the deletion of the levy and the compensation guarantee in the amendment, and the government obtained the bill. A constitutional amendment requires, in each House, a majority of the total membership and two-thirds of the members present and voting. The governing coalition could command the Lok Sabha; it could not command the Rajya Sabha, where the Congress and its sympathisers held enough seats to deny the special majority. The Congress’s demands were specific and were repeated in debate and in its select committee dissent: a ceiling of 18 per cent written into the amendment so that no future Council or government could raise the standard rate beyond it; deletion of the additional one per cent levy, which it characterised as a distortion that would benefit a few manufacturing states at the cost of the common market; the inclusion of petroleum products, alcohol and electricity within the tax; a compensation formula stated in the bill rather than left to subsequent law; and an independent mechanism for settling disputes between the Union and the states. Behind the specific demands lay a credit contest that neither side acknowledged in debate and both conducted in public: the Congress described the tax as its own reform, pilfered and mishandled, while the government described the delay as the obstruction of a measure the opposition had itself introduced in an earlier avatar.

The settlement, when it came in the monsoon session of 2016, conceded the substance of the two demands that went to the integrity of the tax and to the states’ security, declined the rate ceiling, and split the difference on machinery. On 2 August 2016 the government circulated official amendments to the bill. The additional one per cent levy was deleted outright, removing Clause 18 and with it the last origin-based element of the design; the manufacturing states accepted the deletion against the second element of the package. That element was compensation hardened into constitutional text: the amendment itself would oblige Parliament, on the recommendation of the Council, to provide compensation to the states for loss of revenue arising from the introduction of the tax for a period of five years, replacing a tapering executive assurance of the kind Jaitley had earlier described in the House, under which losses would have been met in full for three years and in diminishing proportions thereafter. The amendments further provided that the Council would establish a mechanism to adjudicate disputes between the Union and the states or among states, answering the demand for dispute machinery in the form the states had earlier resisted and then requested, as a mechanism of the Council rather than an authority above it, and that the states’ share of the integrated tax would not form part of the Consolidated Fund of India. The rate ceiling was not conceded. The Council would recommend rates, and no number was written into the Constitution, a refusal the government defended on the ground that a constitutional ceiling would require a further amendment whenever economic conditions required the rate structure to move. The argument over the ceiling is worth recording in its strongest form on both sides, because it is the one point on which the settlement consciously left the taxpayer’s protection to politics. For the ceiling, its advocates argued that a tax levied concurrently by thirty-one governments, at rates recommended by a body whose decisions are in practice unanimous, lacks the ordinary electoral check on rate increases; that a constitutional maximum was the only discipline a future Council could not itself relax; and that the Congress, as the party that had designed the tax, was entitled to insist that the citizen’s exposure be capped at the figure its own design work had contemplated. Against it, the government argued that no rate of any Indian tax has ever been fixed in the Constitution, that rates must move with revenue needs, prices and the structure of the economy, and that a ceiling at eighteen per cent would either bind in a crisis, when a higher rate might be the least damaging expedient, or be amended under pressure and so protect nobody. Both arguments are coherent; the House chose between them on the arithmetic of the special majority, and the question remains available to any future Parliament commanding the votes for an amendment of its own.

On 3 August 2016 the Rajya Sabha debated and passed the bill as amended. The debate ran on the lines the negotiation had laid down. The finance minister presented the official amendments as the completion of the consensus built since 2014 and appealed to the House not to let the remaining difference, the rate ceiling, defeat a reform every party had sponsored in some session. The Congress, speaking through its former finance minister P. Chidambaram among others, supported the bill’s passage while recording that its central demand had not been met and that the discipline of an eighteen per cent ceiling would have to be supplied by politics rather than by the Constitution. Individual members moved amendments to restore the ceiling and on other points; the amendments were pressed and negatived, and the clauses were then adopted. The AIADMK, governing Tamil Nadu and carrying that state’s longstanding objections to the surrender of fiscal autonomy, walked out before the vote; among the members who voted, none voted against, and the amendments the government had circulated the previous day were adopted. The Lok Sabha, which had passed the earlier text, took up the amended bill and passed it on 8 August 2016, completing the parliamentary stage. What had changed between the introduction of the first bill in March 2011 and the passage of the second in August 2016 was not the destination principle, the dual structure or the Council, all of which survived from the original design. What had changed was the distribution of risk. The states had transferred the risk of revenue loss to the Union for five years by constitutional command; the Union had surrendered the additional levy that was its bridge to the manufacturing states and accepted a dispute mechanism it had not originally wanted; and both sides had accepted a Council in which neither could impose its will on the other. The bill that passed was, in that precise sense, the bill the negotiation had been about from the beginning.

Ratification, assent and commencement

Parliamentary passage was the halfway point, not the end. Because the bill altered the Seventh Schedule and the taxing powers of the states, it required ratification by the legislatures of not less than half the states before presentation to the President. The state legislatures moved with a speed that measured how completely the August settlement had done its work. Assam, a consuming state of the north-east whose government had changed hands in the election of that year, ratified first, on 12 August 2016, its assembly approving the amendment unanimously. Bihar followed on 16 August and Jharkhand on 17 August, and through the rest of August the ratifications accumulated: Chhattisgarh, Himachal Pradesh, Gujarat, Delhi, Madhya Pradesh, Nagaland, Haryana, Maharashtra, Mizoram, Sikkim, Telangana and Goa. On 1 September 2016 Odisha became the sixteenth legislature to ratify, carrying the amendment past the constitutional threshold of one-half. Ratifications continued to arrive after the requirement was met, from Rajasthan and Puducherry in the days around the assent to, in the following year, states whose assemblies had not been in session or whose politics required more time; Andhra Pradesh ratified on 8 September 2016, the day of the assent itself, and thereafter Punjab, Tripura and others through the autumn. In the end the amendment was ratified by every state legislature, a completeness the constitutional requirement does not demand and the politics of 2011 would not have predicted: Tamil Nadu ratified in June 2017, Kerala the same month, Jammu and Kashmir on 5 July 2017, and West Bengal, the state whose finance minister had chaired the Empowered Committee for its first decade, in August 2017. The late ratifications carried no legal necessity, the threshold having been crossed eleven months earlier, and precisely for that reason they are evidence of what the settlement had achieved: governments that had not needed to sign, signed, because the text in force was one their objections had already improved. The President, Pranab Mukherjee, who as finance minister had introduced the first bill in 2011, gave his assent on 8 September 2016, and the measure was notified the same day as the Constitution (One Hundred and First Amendment) Act, 2016, the article at that link being the owner of the amendment’s text and of the Council’s subsequent career. Article 279A, creating the Council, was brought into force on 12 September 2016, and the remaining provisions of the amendment followed on 16 September 2016.

What the ratified text did, stated at the level a passage history requires, can be summarised as the embodiment of the settlement in five moves. It inserted Article 246A, conferring on Parliament and on every state legislature the concurrent power to legislate on the goods and services tax, with Parliament’s power exclusive for inter-state supplies. It inserted Article 269A, under which the integrated tax on inter-state supplies is levied and collected by the Union and apportioned between the Union and the states in the manner Parliament provides on the Council’s recommendation, with imports treated as inter-state supplies. It defined the tax in Article 366 so as to exclude alcoholic liquor for human consumption, and it gave the five petroleum products their deferred position through the Council’s power under Article 279A to recommend the date of their levy. It constituted the Council in Article 279A, with the composition and the weighted voting the negotiation had settled, and charged it with recommending rates, thresholds, exemptions, the model laws and the principles of place of supply, and with establishing the dispute mechanism. And in its eighteenth section it directed Parliament to provide, by law on the Council’s recommendation, compensation to the states for the loss of revenue arising from the introduction of the tax, for five years. Around those moves, the Act recast the tax entries of the Seventh Schedule, amended the Sixth Schedule for the autonomous councils, and repealed the constitutional basis of the levies the new tax would subsume. The detail of the enacted provisions, and the Supreme Court’s later reading of the Council’s recommendatory character, belong to the article on the amendment itself; the point for the passage record is that every operative feature of the text answers to a named stage of the negotiation set out in this article.

The sequence deserves a moment’s attention, because it reversed the pattern of the preceding five years. Between 2011 and 2016 the states had been the brake on the reform; between August and September 2016 they were its accelerator, ratifying in three weeks a constitutional change they had resisted in draft for half a decade. The difference was not a change of heart about the destination principle, whose incidence they understood as well in 2016 as in 2011. The difference was the text they were asked to ratify. The compensation obligation stood in the amendment itself. The additional levy was gone. The Council’s dispute mechanism was theirs. Ratification politics is ordinarily the slowest stage of an Indian constitutional amendment; for this amendment it was the fastest, because the bill that reached the assemblies had already been negotiated with the governments that controlled them.

The gap between the introduced bill and the enacted amendment

The analytical heart of any passage history is the distance between the text first presented and the text finally enacted, and for this reform the distance can be measured against two baselines: the bill of 2011, and the bill of 2014 as introduced. Against the bill of 2011, the enacted amendment differs at every point where the first negotiation had failed. The Dispute Settlement Authority, a free-standing body interposed between the governments, appears in the bill of 2011 as a central feature and in the enacted text not at all; its function survives only as a direction that the Council establish a mechanism for adjudicating disputes, which is to say the governments kept the dispute function inside the room where they sit. The treatment of petroleum inverted: the bill of 2011 excluded the petroleum products from the tax, while the enacted text brought them within the definition of goods and reserved their levy to a date the Council would recommend, converting an exclusion into a deferral and giving the states a collective, rather than individual, grip on the timing. The compensation promise travelled from the political assurances surrounding the bill of 2011, through a permissive version in the bill of 2014, to the enacted direction that Parliament provide compensation for five years. The Council itself survived both drafts, but its enacted form fixed the voting weights, one-third for the Union and two-thirds for the states collectively with a three-fourths threshold for decisions, in terms that answered the earliest of the state objections, the fear of being outvoted in a body chaired by the Union finance minister.

Against the bill of 2014 as introduced, the changes of August 2016 are fewer and sharper, and they are the changes the Rajya Sabha’s arithmetic extracted. Clause 18, the additional levy, was deleted. The compensation clause was hardened from an assurance of principle into the five-year direction. The Council was charged with the dispute mechanism. The states’ share of the integrated tax was placed outside the Consolidated Fund of India, so that the apportionment to the states would not depend on the annual appropriations process. Everything else in the introduced text, the concurrent power, the integrated levy, the Council’s composition and recommendatory charter, the deferred petroleum levy and the exclusion of alcohol, passed into the Act substantially as drafted in December 2014, which is itself a finding: the drafting of the second bill, built on the standing committee’s report, the Empowered Committee’s recommendations and three years of revision, survived Parliament with its architecture intact, and the parliamentary stage changed only the terms of the bargain, not the design of the tax.

A small formal point completes the comparison and is a standing source of examination error. The measure was introduced as the Constitution (One Hundred and Twenty-second Amendment) Bill because amendment bills are numbered in the order of their introduction in a sessional sequence, and it was enacted as the Constitution (One Hundred and First Amendment) Act because amendment Acts are numbered in the order in which they receive assent and take their place in the Constitution. The gap between the two numbers is the count of amendment bills introduced but not enacted in the interval, the lapsed bill of 2011 among their predecessors. The legislative history is partly written in that arithmetic.

What each concession bought

The enacted settlement rested on five concessions, each identifiable in a text or a drafting change, each with a price. Read together, they explain both why the amendment passed in 2016 and why the tax that followed took the shape it did. The compensation guarantee was the keystone. The states, asked to surrender the power to set the rates of their largest taxes, demanded and received a constitutional direction that Parliament provide compensation for the revenue loss of the first five years, given effect by the compensation statute’s formula of the base year 2015-16 grown at fourteen per cent a year, a growth assumption generous by the standards of the states’ own recent collections. The fiscal argument that the guarantee set running, about what would happen when rates, collections or the economy fell short of the assumed path, is the subject of the article on GST and Centre-state federalism, which owns the compensation dispute and its aftermath.

Concession	What the states demanded	What the Union conceded	What the concession cost or foreclosed
The compensation guarantee	Insurance against revenue loss, in the constitutional text rather than in a ministerial assurance	A direction in the amendment itself that Parliament compensate states for five years, implemented on a base of 2015-16 revenues grown at fourteen per cent a year	A large and rigid Union liability, whose cost depended on collections the Union did not control, and a five-year horizon after which the states’ exposure returned in full
The exclusion of specified products	Retention of alcohol entirely, and deferral of petroleum, the two richest state tax handles	Alcohol for human consumption placed outside the definition of the tax; five petroleum products brought within the frame but taxed only from a date the Council recommends	A permanent hole in the base at alcohol and a long-deferred one at petroleum; input tax credit chains broken wherever fuel enters production or transport
The dropping of the additional one per cent levy	A transitional origin-based levy for manufacturing states, levied by the Union and assigned to the state of origin	The levy was carried in the bill as introduced and through the select committee, then deleted by the government’s own amendments in August 2016	The manufacturing states lost their bridge payment and relied entirely on compensation; the design gained coherence, since an origin levy contradicts the destination principle
The voting weights in the Council	Protection against being outvoted by the Union, and by each other	The Union’s vote weighted at one-third of the total, the states’ votes together at two-thirds, and decisions to require three-fourths of the weighted votes cast	Neither side can carry a decision alone; the Union holds a blocking share, the states hold a collective blocking share, and rate changes require a breadth of agreement that can delay them
Retention of state power over the excluded goods	Continued exclusive taxation of the goods left outside the common tax	States kept their excise on alcohol, their sales tax and VAT on petroleum products pending the Council’s decision, and levies such as electricity duty outside the tax	A dual fiscal world persists at the boundary of the tax, with the excluded sectors administered under the old law and the old disputes, and a standing incentive to postpone inclusion

Each row records a transfer of risk or of revenue, and each was rational from the seat of the party that demanded it. The compensation guarantee transferred the transition risk to the Union, which was the party best able to bear it and the party that wanted the reform most; its cost was a contingent liability that later became a live dispute when collections faltered. The product exclusions kept the states solvent and kept the Union’s own excise base partially intact, at the price of the credit chain, which is the tax’s reason for existing, breaking down in the two sectors where fuel and liquor enter the cost of nearly everything else. The deletion of the additional levy was the one concession that improved the design while costing a constituency, and it was purchasable only because the compensation guarantee had been hardened in the same week. The voting weights made the Council a body neither side could capture, which protected the states against the Union and the Union against a coalition of states, and which guarantees that the hardest decisions of the tax require something close to the consensus that produced the tax itself. The retention of state power over the excluded goods converted a surrender into a partition: the states gave up the general power to tax goods and services and kept the particular powers that mattered most to their treasuries.

Two boundary cases complete the map of the settlement and show its logic at work. Electricity was left outside the tax altogether, the states retaining their duty on the consumption of electricity and the Union its position, because electricity is an input into everything and a political price in every state, and neither side would pool it in the first settlement; the cost, a further break in the credit chain for every business that buys power, was accepted as the price of leaving the question for another negotiation. Tobacco travelled the opposite way: it was brought within the tax while the Union retained its power to levy excise duty on tobacco products in addition, a deliberate double levy unique in the design, preserved because the Union would not surrender its most elastic sin revenue and the states, who had never taxed tobacco manufacture, had nothing to trade against the demand. Read together, the two cases state the settlement’s rule. A product stayed out where a government that could block the amendment drew essential revenue from taxing it alone; a product came in, on special terms, where the revenue at stake belonged to the government that needed the amendment most. Principle supplied the vocabulary of the negotiation. Revenue supplied its outcomes.

A reading of the passage record that treats these as defects introduced by politics misses their function. A constitutionally entrenched negotiation among twenty-nine governments does not produce the design a single draftsman would choose. It produces the design that can be ratified. The concessions were not decorations on the settlement; they were the settlement, and the enacted text is best read as their sum.

From amendment to launch: the Council and the statutes of 2017

The amendment conferred power; it levied nothing. The ten months between assent and launch were spent building the tax inside the constitutional frame, and they were spent in the institution the amendment created. The machinery had been prepared in advance of the text. A draft model goods and services tax law had been placed in the public domain in June 2016, while the amendment was still before the Rajya Sabha, so that the Council, when constituted, inherited a working draft rather than a blank page; the Cabinet approved the Council’s creation and its secretariat on 12 September 2016, the day Article 279A came into force, with the Union revenue secretary as the Council’s ex-officio secretary. The President constituted the Goods and Services Tax Council on 15 September 2016, with the Union finance minister in the chair, the Union minister of state for revenue as a member, and the finance or taxation minister of every state as members. The Council held its first meeting on 22 and 23 September 2016 and set to work on the agenda the amendment had assigned it: the model laws, the rates and thresholds, the place of supply principles, the treatment of the area-based exemptions inherited from the excise regime, and the administrative division of taxpayers between the Union and the states. At that first meeting it fixed the exemption threshold at twenty lakh rupees of turnover for most states and ten lakh for the north-eastern and hill states, and settled the broad division of administrative control, and by the second meeting it had confirmed 2015-16 as the compensation base year with the fourteen per cent growth assumption. The administrative division deserves a sentence of its own, because it closed one of the oldest arguments in the design. A dual tax assessed by two administrations threatened every dealer with two audits, two assessments and two appellate paths, and the states and the Union had contested the roster since the Joint Working Group’s report. The Council’s settlement gave the states sole administrative charge of taxpayers up to a turnover of one and a half crore rupees, divided the larger taxpayers between the two administrations, and kept the existing service tax assessees with the Union for the transition. The settlement was a concession of the same family as the others: administrative territory, like revenue, was partitioned by negotiation rather than assigned by principle, and the partition lines chosen in September 2016 were the ones both administrations could staff. Over the following six months the Council met twelve times and approved the draft statutes clause by clause, so that the laws Parliament enacted in the spring of 2017 were, in their substance, the Council’s laws, voted on by the state legislatures’ own ministers before any assembly saw them.

How many separate laws did the launch of GST require?

Four central statutes and a state statute in every state. Parliament enacted the Central, Integrated, Union Territory and Compensation to States Acts in April 2017, and each state legislature enacted its own State Goods and Services Tax Act on the Council’s model before the appointed day.

The Council’s autumn meetings also fixed the shape of the rate structure, settling in November 2016 on four principal rates, with a cess on luxury and demerit goods whose proceeds would feed the compensation fund, and rejecting the single-rate design the Kelkar reports had contemplated as the price of a structure every state could sign. The same meetings had to settle the inherited geography of the old system: the area-based excise exemptions by which the Union had drawn industry to the hill and north-eastern states could not survive a destination-based credit tax in their original form, and the Council agreed that the exemptions would be honoured by collecting the tax and reimbursing the exempted units from the budget, a conversion of a tax concession into an expenditure that preserved the promise while submitting it, for the first time, to annual appropriation. The episode is minor in the launch record and characteristic of the whole: where a commitment of the old regime could not be carried into the new one, the settlement bought it out rather than repudiating it, and charged the cost to the party that had made the commitment. The fitment of thousands of goods and services into those rates occupied the Council through the spring of 2017 and continued past the launch, a reminder that the amendment had settled the architecture and deliberately left the arithmetic to a standing negotiation. The government’s original ambition of an April 2017 start gave way, as the Council’s legislative programme lengthened, to the appointed day of 1 July 2017.

The four central bills, approved by the Union Cabinet on 20 March 2017, were passed by the Lok Sabha on 29 March 2017 and by the Rajya Sabha on 6 April 2017, the upper House returning them without amendment as money bills, and they received presidential assent on 12 April 2017, the compensation statute becoming Act 15 of 2017. The working detail of the four statutes and of the state laws that mirror them belongs to the article on the GST Acts of 2017. The state legislatures then enacted their State Goods and Services Tax Acts through April, May and June 2017, in special sessions in several states, each assembly in substance re-enacting the model its own finance minister had already approved in the Council. Jammu and Kashmir came last, and its route was constitutionally distinct. Under the state’s special position, the amendment’s application to Jammu and Kashmir required the state legislature’s own concurrence, and the question was debated there as a question about the state’s fiscal autonomy in its sharpest form, with the state’s parties divided on whether entry into the common tax surrendered a protection the state’s constitutional position existed to preserve. The legislature ratified the amendment on 5 July 2017 and completed its state law within the same week, and the tax applied there from 8 July 2017, a week behind the rest of the country. The delay cost the state’s trade a week of uncertainty at the border; the sequence confirmed, in the last instance as in the first, that in this reform the consent of each constituent government was a legal fact and not a courtesy.

The state enactments of the spring of 2017 drew a criticism the passage record answers in advance: that assemblies passed, in sittings of a day or two, model laws of great length and complexity without scrutiny. The criticism mistakes where the scrutiny had occurred. The text each assembly enacted had been approved clause by clause in the Council by that state’s own finance minister, on the advice of that state’s own officers, in a process that ran for six months and twelve meetings; the assemblies were not being asked to examine a stranger’s bill but to ratify their own government’s drafting, and the brevity of the enactment stage is the measure of how completely the legislative work had migrated, for this reform, from the floor of the houses to the committee room of the federation. Whether that migration is a precedent to celebrate or to guard against is a genuine question for the student of Parliament, and it is the question the passage of the goods and services tax leaves behind it for every future reform of its scale.

What happened at the midnight sitting that launched GST?

Parliament met in a sitting in the Central Hall on the night of 30 June and 1 July 2017, at which the President and the Prime Minister addressed the members and the tax was launched at midnight. The sitting enacted nothing; the law stood commenced by notification, and several opposition parties, the Congress among them, stayed away from the ceremony.

The launch closed the passage story with a symmetry its participants noticed. The President who addressed the midnight sitting was Pranab Mukherjee, who had introduced the first bill six years earlier. The Prime Minister who addressed it was Narendra Modi, whose state government had been among the first bill’s most effective opponents. The tax that came into force at that midnight was neither the bill of 2011 nor the bill of 2014 as introduced. It was the bill of August 2016, the negotiated text, carrying its five concessions into force with it.

What the passage record predicts

A legislative history earns its place by what it explains about the law’s subsequent behaviour, and the GST passage record predicts three durable features of the tax’s career. The first is the Council’s method. An institution designed so that neither the Union nor the states can be outvoted, and preceded by a decade in which a committee of state finance ministers designed the tax by consensus before any institution existed, will default to consensus and will treat voting as failure. The later judicial reading of the Council’s recommendatory character, under which its recommendations do not bind the Union or the states, surprised only those who had not followed the negotiation, in which the recommendatory form was itself one of the concessions: the states accepted a joint institution on the express footing that it would recommend and they would legislate. The second prediction concerns the compensation horizon. A guarantee with a five-year term, sold to the states as the price of surrender, created a date on which a large fiscal argument was certain to occur, and the record of 2011 to 2014, in which the states bargained hardest precisely over insurance, predicted both that the argument would occur and that it would be conducted in the language of broken promise. The third concerns the exclusions. Petroleum and alcohol were not oversights awaiting tidying; they were retained revenue, held by named governments for stated reasons, and their eventual inclusion will require a second negotiation of the same character as the first, in which the question will again be what the states are paid, and in what instrument, for letting go. A fourth prediction follows from the method rather than from any clause. The passage demonstrated a working technology for federal fiscal reform in India, committee negotiation among the affected governments before drafting, insurance drafted as debt, exclusions priced rather than condemned, and a standing joint institution with weighted votes to hold the settlement afterwards. That technology is available to the next reformer of a divided fiscal field, in electricity, in land, in the municipal finances, and the passage record of the goods and services tax will be its manual and its warning in equal measure, since it shows both that the method works and what it costs in time: the better part of two decades, and a design that bears the marks of every government that had to be persuaded.

The passage record also disposes of the most durable popular criticism of the launch, that the tax was rushed. Against the constitutional negotiation, the charge fails on the dates: seventeen years from proposal to launch, two amendment bills, two parliamentary committee examinations, a decade of design work in the Empowered Committee, and a ratification campaign. Against the implementation timetable, the charge is accurate and should be conceded precisely. Between the Council’s constitution in September 2016 and the appointed day in July 2017, the system had to complete the model laws, four central statutes and twenty-nine state statutes, a rate fitment for the whole tariff, the migration of tens of lakhs of existing taxpayers onto a new technology platform, and the drafting of the rules and forms under which returns would be filed; the original target of April 2017 was abandoned for July, and elements of the return machinery were simplified and deferred after launch when the full design proved more than the platform and the taxpayer base could absorb at once. The two findings are consistent, and the history explains both. The negotiation could not be hurried because its parties held vetoes; the implementation could not be extended because the amendment’s own logic, and the fiscal year, set the date, and the parties who had spent seventeen years agreeing the tax were not willing to spend a further year preparing it. The rush was the negotiation’s final cost, transferred from the legislatures to the administrators.

For the researcher assembling the record of this negotiation, the primary materials are scattered by design: two bills and a lapsed bill’s committee report, a discussion paper issued by a committee of state governments, a select committee report with its dissent, the debates of three parliamentary sessions and the ratification resolutions of thirty legislatures. Keeping that file coherent across years of argument is the same species of problem faced by anyone holding sensitive negotiating material, and the discipline described in the account of how journalists use VaultBook to protect sources applies to it directly: one private store, controlled access, and a record that survives the argument it documents.

The credit question, settled by the sequence

Public argument about the reform habitually assigns it an author, and the passage record refuses every single name offered. The proposal was written by a task force appointed by a government of the National Democratic Alliance under Prime Minister Atal Bihari Vajpayee. The reform was announced with a target date by the finance minister of a United Progressive Alliance government, designed in a committee of state finance ministers chaired for a decade by a finance minister of the Left Front in West Bengal and for two years by a finance minister of the BJP in Bihar, and first introduced as a bill by a Congress finance minister, Pranab Mukherjee, who as President gave the successful bill his assent. The successful bill was introduced and carried by a BJP finance minister, Arun Jaitley, under a prime minister who had opposed its predecessor as a chief minister; its committee scrutiny in the second attempt was conducted by a select committee of the Rajya Sabha whose work the Cabinet largely accepted; and its final concessions were extracted by the Congress in the exercise of a blocking minority it had not possessed when it held the finance ministry. Jaitley, winding up the debate on the four statutes in the Rajya Sabha in April 2017, said of the tax that successive governments had contributed to it and that it was a collective property, and the passage record bears him out in detail. The accurate attribution is institutional: the reform was made by the committee system of Indian fiscal federalism, the Empowered Committee first and the Council after it, operating across four changes of government in Delhi and dozens in the states. Single-party accounts of its authorship are not simplifications of the record. They are contradictions of it.

The Council’s constitutional existence also closed the Empowered Committee’s. The committee of state finance ministers that had designed first VAT and then the GST design had no function once its members sat, with weighted votes and a constitutional mandate, in the Council, and its former chairman Sushil Kumar Modi remarked in 2018 that the Empowered Committee had outlived its purpose and lost relevance with the Council’s formation. The observation, offered as a gibe in a political argument, is the correct institutional epitaph. The Empowered Committee was the Council in rehearsal: the same ministers, the same officials, the same consensus method, differing only in that its agreements bound nobody until a bill embodied them. The seventeen years were, in substantial part, the time taken to move the negotiation from a room whose conclusions were advisory into a room whose conclusions were the law’s raw material, and to persuade the occupants of both rooms that the move was safe.

The ledger’s lessons for the next federal bargain

The passage record is a usable manual for any future reform that requires the states to surrender a tax, and its lessons are specific enough to state. The first is that the negotiation must precede the bill. Every attempt to draft first and consult afterwards, from the Union’s draft amendments of 2010 to the timetable of 2010 itself, failed or stalled; every stage that succeeded, the Joint Working Group’s design, the Empowered Committee’s revisions of the pending bill, the select committee’s consultation, the Council’s clause-by-clause approval of the statutes, put the governments that had to consent in the room before the text was fixed. The second is that compensation works only when it is drafted like a debt: a named base, a named growth rate, a named term, and a constitutional or statutory obligation that survives the minister who offered it. The tapering assurances of the central sales tax phase-out and the early GST years bought nothing because they were priced by their recipients as revocable, and they were. The third is that carve-outs are cheaper than vetoes. The Union obtained its amendment by letting the states keep alcohol outright and petroleum on a deferred fuse, at a permanent cost to the tax’s design; a reformer who insists on a complete base in a federation of vetoes will obtain, as the reformers of 2011 obtained, a complete base in a bill that lapses. The fourth is that the institution outlasts the bargain. The weighted voting of the Council, derided at the drafting stage as a recipe for deadlock, is what made the surrender of rate-setting power acceptable to the surrendering parties, because it guaranteed that the power, once pooled, could not be used against any of them by the others. A future negotiator reading this record for method will find its order instructive in itself: design in committee, agree the insurance, price the exclusions, build the joint institution, and only then introduce the bill. The order is not cosmetic. Each stage of the GST passage that observed it advanced, and each stage that inverted it, announcing a date before the design or a design before the insurance, produced a missed deadline and a harder negotiation afterwards, until the sequence was finally followed to the end.

The verdict

GST took seventeen years because it was, from the first Kelkar report to the last ratification, a redistribution of revenue between governments that each held a veto over some stage of its passage, and vetoes are retired by compensation, not by persuasion. The standard account, in which a good reform was delayed by obstruction, fails in both directions: the states that resisted were defending revenues the reform genuinely took from them, and the Union that finally conceded the compensation guarantee was not being generous but paying the market price for a constitutional amendment it could not pass or ratify alone. What the passage record shows, when it is read as a negotiation rather than as a chronicle of delay, is a federation learning the cost of its largest fiscal decision and itemising the bill in the text of the amendment itself. The tax India received in July 2017 was the tax that could be bought. The five concessions were the purchase price, and their consequences, in the credit chain, in the Council’s arithmetic and in the compensation dispute that followed, are the subsequent history of the tax working itself out. That is the sense in which the legislative history of the goods and services tax is not the preface to the law. It is the law’s first chapter.

Frequently Asked Questions

Q: When was GST first proposed in India?

The proposal entered the official record in 2002, when the Task Force on Direct and Indirect Taxes chaired by Vijay Kelkar recommended a comprehensive goods and services tax to replace the fragmented system of excise, service tax and state sales taxes. A second Kelkar body, the task force on implementing the fiscal responsibility legislation, repeated the recommendation in 2004, and the Thirteenth Finance Commission, also chaired by Kelkar, proposed in 2009 that the country move to the new tax backed by a compensation grant. The Union Budget for 2006-07 converted the expert proposal into a government commitment by announcing that a goods and services tax would be introduced from 1 April 2010.

Q: What did the Kelkar Task Force recommend about GST?

The Task Force on Direct and Indirect Taxes, appointed in 2002 under Vijay Kelkar, recommended replacing the patchwork of central excise, service tax and state sales taxes with a comprehensive value added tax on goods and services, levied across the Union and the states, with input tax credit flowing through the production and distribution chain so that the tax would fall on final consumption rather than accumulating at each stage. The fiscal responsibility task force of 2004 treated the same reform as part of fiscal consolidation, and the Thirteenth Finance Commission’s report of 2009 sketched a model version of the tax at a single combined rate, supported by a grant of fifty thousand crore rupees to bring the states in.

Q: Why did the 115th Amendment Bill on GST lapse?

The Constitution (One Hundred and Fifteenth Amendment) Bill, 2011 was introduced in the Lok Sabha on 22 March 2011 and referred to the Standing Committee on Finance, which reported in August 2013. The bill was never taken up for consideration and passing. Negotiations between the Union and the Empowered Committee over the Council’s design, the proposed Dispute Settlement Authority and compensation continued into the spring of 2014 without a settled text, and when the fifteenth Lok Sabha was dissolved in May 2014, every bill pending in that House that had not been passed lapsed with it, the GST bill among them. The lapse was procedural in form; its cause was the absence of the federal agreement the bill needed.

Q: What was the Empowered Committee of State Finance Ministers?

The Empowered Committee was a standing body of the finance ministers of the states, created after a conference of chief ministers and state finance ministers in November 1999 to design the state-level value added tax, and later charged with preparing the road map for the goods and services tax. It had no statutory basis and its conclusions bound nobody, but because every state government sat in it, its positions carried the weight of a collective state view. Chaired by Asim Kumar Dasgupta of West Bengal from its formation until 2011, and then by Sushil Kumar Modi of Bihar until 2013, it produced the White Paper on state VAT, the Joint Working Group’s dual GST design and the First Discussion Paper of 2009, and it negotiated the states’ terms with the Union until the GST Council replaced it in 2016.

Q: Why was the additional one per cent levy dropped from the GST bill?

The Constitution (122nd Amendment) Bill as introduced in 2014 carried, in its Clause 18, an additional tax of up to one per cent on inter-state supplies of goods, to be collected by the Union and assigned to the state of origin, as a transitional comfort for manufacturing states that stood to lose under a destination-based tax. The Congress demanded its deletion as a distortion that contradicted the destination principle, and producing states defended it as their bridge out of origin-based revenues. In the settlement of August 2016 the government deleted the levy by its own official amendments, because the simultaneous hardening of the compensation guarantee, five years of compensation written into the amendment itself, gave the manufacturing states the security the levy had been meant to provide.

Q: How long did GST take from proposal to rollout?

Seventeen years separate the Kelkar Task Force’s recommendation of 2002 from the launch of the tax on 1 July 2017, and the interval divides into recognisable phases. Design occupied the years to 2009, ending with the Empowered Committee’s First Discussion Paper. The first amendment bill occupied 2011 to 2014 and lapsed. The second bill occupied December 2014 to August 2016, when it passed, and ratification, the Council’s constitution and the enactment of the four central statutes and the state laws occupied the ten months to the launch. The frequently missed target dates, April 2010 and April 2016 among them, mark the points at which announced timetables ran ahead of the federal bargain the reform required.

Q: Which states resisted GST and why?

Resistance followed economic structure more than party. Manufacturing states, Gujarat, Maharashtra and Tamil Nadu foremost, resisted longest, because a destination-based tax moves revenue from the state where goods are produced to the state where they are consumed, and because the abolition of the central sales tax removed an origin-based levy they collected. Gujarat’s governments pressed the fiscal autonomy objection, the unrepaired losses of the central sales tax phase-out and the unreadiness of the information technology backbone; Tamil Nadu under J. Jayalalithaa pressed autonomy and producing-state losses, and its party walked out of the Rajya Sabha in 2016 rather than vote for the amendment. Consuming states of the north and east stood to gain from the destination principle and settled earlier, once compensation for transitional losses was guaranteed.

Q: Was the GST bill referred to a select committee?

Yes, in its second attempt. The Constitution (122nd Amendment) Bill, 2014 was passed by the Lok Sabha on 6 May 2015 and, on reaching the Rajya Sabha, was referred on 14 May 2015 to a select committee of twenty-one members chaired by Bhupender Yadav. The committee consulted state governments, experts and stakeholders and reported on 22 July 2015, broadly endorsing the bill, retaining the additional one per cent levy with safeguards and recommending a firmer compensation arrangement, while Congress members recorded their dissent on the rate ceiling and other demands. The Union Cabinet accepted the greater part of the recommendations within the week. The first bill of 2011 had gone instead to the Standing Committee on Finance, the ordinary committee route for a bill introduced in the Lok Sabha.

Q: What did the Empowered Committee’s First Discussion Paper of 2009 propose?

Released in New Delhi on 10 November 2009 under Asim Kumar Dasgupta’s chairmanship, the First Discussion Paper set out the proposed structure of the tax in public for the first time. It proposed a dual goods and services tax, a central component and a state component levied concurrently on the same base, with an integrated tax on inter-state supplies; the subsumption of central excise, service tax and state VAT, central sales tax, entry tax and allied levies; zero rating of exports and taxation of imports; and the constitutional amendments needed to confer concurrent taxing power. It deliberately left open the rates, the exemption threshold, the treatment of petroleum products and alcohol, and the compensation of states that would lose revenue, the four questions the following seven years of negotiation were spent answering.

Q: What was the GST Dispute Settlement Authority proposed in the 2011 bill, and why was it dropped?

The Constitution (115th Amendment) Bill, 2011 proposed, alongside the GST Council, a Goods and Services Tax Dispute Settlement Authority empowered to adjudicate disputes between the Union and the states, or among states, arising where a government departed from the Council’s recommendations and another government’s revenue suffered. The states read the Authority as a tribunal placed above the negotiation, created under Union sponsorship and competent to sit in judgment on a state legislature’s fiscal choices, and it became one of the principal institutional objections to the bill. The Standing Committee on Finance recorded the concern in its 2013 report. The 2014 bill omitted the Authority, and the settlement of 2016 placed the dispute function inside the Council itself, directing the Council to establish a mechanism for adjudicating such disputes.

Q: What did the Standing Committee on Finance recommend on the 2011 GST bill?

The Standing Committee on Finance, chaired by Yashwant Sinha, examined the Constitution (115th Amendment) Bill for over two years, took evidence from the finance ministry, the Empowered Committee’s leadership, industry and experts, and presented its seventy-third report on 7 August 2013. The report supported the introduction of the goods and services tax and recommended that the bill be taken forward, while recording the states’ anxieties in detail: the effect on fiscal autonomy, the adequacy and form of compensation for revenue loss, the composition and voting arrangements of the proposed Council, and the design of the proposed Dispute Settlement Authority, on which it urged reconsideration. The report became part of the working file from which the successor bill of 2014 was drafted.

Q: How did the 2014 GST bill differ from the bill that had lapsed in 2011?

Four differences mark the negotiation’s movement. The Dispute Settlement Authority of the 2011 bill disappeared, leaving disputes to the Council. Petroleum products, excluded from the tax in 2011, were brought within its constitutional frame in 2014 but with their levy deferred to a date the GST Council would recommend. The 2014 bill added a transitional device the earlier bill lacked, the additional tax of up to one per cent on inter-state supplies of goods for the manufacturing states, later deleted in 2016. And the 2014 bill was drafted, from the Empowered Committee’s recommendations and the Standing Committee’s report upward, with the states inside the drafting process, where the 2011 bill had been presented to them and defended. Alcohol for human consumption remained outside the tax in both.

Q: What did the Rajya Sabha Select Committee recommend on the GST bill in 2015?

The select committee of the Rajya Sabha, twenty-one members under Bhupender Yadav, reported on 22 July 2015 after consulting state governments, experts and stakeholders. It broadly endorsed the Constitution (122nd Amendment) Bill as passed by the Lok Sabha, recommended retaining the additional one per cent levy on inter-state supplies for an initial period with safeguards for the common market, proposed that compensation to the states be placed on a firmer footing through a dedicated compensation fund, and addressed the position of local bodies whose levies the tax would subsume. Congress members dissented, pressing for a ceiling of eighteen per cent in the constitutional text, the deletion of the levy and the inclusion of petroleum products and alcohol. The Cabinet accepted the greater part of the committee’s recommendations on 29 July 2015.

Q: Why did the GST bill remain stalled in the Rajya Sabha through 2015 and into 2016?

A constitutional amendment needs, in each House, a majority of total membership and two-thirds of members present and voting. The governing coalition commanded the Lok Sabha but not the Rajya Sabha, where the Congress and its sympathisers could deny the special majority. The Congress’s stated terms were a ceiling of eighteen per cent in the amendment, deletion of the additional one per cent levy, inclusion of petroleum products and alcohol, a compensation formula in the text and an independent dispute mechanism. The government worked through 2015 to assemble support among regional parties, whose state governments’ interests were largely met by the bill’s existing terms, which progressively isolated the Congress position. The settlement came only when the government conceded the levy and the compensation guarantee in August 2016.

Q: Which state ratified the GST amendment first, and how quickly was the required majority reached?

Assam ratified first, its legislative assembly approving the amendment unanimously on 12 August 2016, four days after the Lok Sabha completed the parliamentary stage. Because the amendment altered the Seventh Schedule, it required ratification by the legislatures of not less than half the states under the proviso to Article 368. Bihar followed on 16 August and Jharkhand on 17 August, and the ratifications accumulated through the month, Gujarat, Maharashtra, Haryana, Delhi, Madhya Pradesh and Goa among them. Odisha became the sixteenth legislature to ratify on 1 September 2016, crossing the constitutional threshold in three weeks. President Pranab Mukherjee gave assent on 8 September 2016, and every state legislature ratified in the end, the last of them in 2017.

Q: Why did the states insist on keeping petroleum products and alcohol out of the GST net?

Because the two were among the largest revenue handles the states possessed and the ones whose rates they most valued controlling. State taxes on petroleum products, levied through sales tax and VAT outside the credit chain, and state excise on alcohol, together supplied a very large share of the states’ own tax revenue, and both could be adjusted quickly when a treasury was under pressure. Pooling them in a common tax at Council-recommended rates would have surrendered that discretion for an uncertain share of a common pool. The settlement split the difference: alcohol for human consumption was excluded from the tax altogether, while the five petroleum products were brought within the constitutional definition and their levy deferred until the GST Council recommends a date.

Q: What did the long dispute over Central Sales Tax compensation do to the GST negotiations?

The central sales tax, an origin-based levy on inter-state sales, was cut from four per cent to three per cent in 2007 and to two per cent in 2008, in agreed anticipation of GST, with the states to be compensated for the phase-out on a declining schedule. When GST failed to arrive on the assumed date of April 2010, the compensation question outlived its schedule, and disputes over computation, duration and arrears hardened into a settled state conviction that Union compensation promises expire before the losses do. The GST negotiations inherited that conviction whole. It is the direct ancestor of the states’ central demand of 2016: compensation for five years, in the constitutional text, on a fixed base year and a fixed growth rate, rather than in ministerial assurance.

Q: How did the positions of the Congress and the BJP on GST change between 2011 and 2016?

The two parties exchanged positions almost exactly. Under the United Progressive Alliance, Congress finance ministers set the target date and introduced the 2011 bill, while BJP-ruled states, Gujarat and Madhya Pradesh in the lead, supplied the organised resistance on fiscal autonomy, producing-state losses and unready technology. After 2014 the BJP in Union government introduced the successor bill within seven months and staked its programme on passing it, under a prime minister who as Gujarat’s chief minister had resisted the predecessor; and the Congress, holding the balance in the Rajya Sabha, resisted through 2015 and into 2016, demanding a constitutional rate ceiling, deletion of the additional levy and firmer compensation. The arguments on both sides remained constant throughout. The seats from which they were argued had changed.

Q: Why was the original target of launching GST on 1 April 2010 missed?

The date was announced in the Union Budget for 2006-07, before the design existed, and the design alone consumed the years to the First Discussion Paper of November 2009. Behind the design stood the constitutional obstacle: the tax entries of the Seventh Schedule divided the field between the Union and the states, so nothing could be levied until a constitutional amendment conferred concurrent power, and no amendment bill had even been introduced when the date passed. Ratification by half the state legislatures would then have been needed, on a design the states had not accepted. The Empowered Committee’s work programme, the Joint Working Group’s report of 2007 and the discussion paper of 2009, assumed the amendment would lead and the launch would follow; the budget’s date assumed the reverse, and the sequence proved impossible.

Q: What happened between the passage of the 101st Amendment and the launch of GST in July 2017?

The amendment conferred power but levied nothing, and the following ten months built the tax. The President constituted the GST Council on 15 September 2016; the amendment’s main provisions came into force on 16 September; and the Council, meeting first on 22 and 23 September 2016, fixed thresholds, confirmed the compensation base year of 2015-16 with a fourteen per cent growth assumption, divided administrative control between the Union and the states, settled a four-rate structure with a cess for the compensation fund, and approved the draft statutes clause by clause over twelve meetings. Parliament passed the four central Acts in the spring of 2017, with presidential assent on 12 April, every state legislature enacted its State GST Act by the end of June, Jammu and Kashmir completing its own route a week later, and the tax came into force at the midnight sitting of Parliament on 1 July 2017.