Friday, 30 June 2017. One minute to midnight.
Parliament’s Central Hall is holding a midnight session to launch a tax that replaces seventeen older levies. The roll-call of the opposition is split. The JD(U), the NCP, the BJD, the Samajwadi Party and the JD(S) have come. The Congress, the Trinamool Congress, the Left, the BSP, the DMK and the RJD have boycotted the ceremony. The Samajwadi Party’s Ram Gopal Yadav gives his reason for attending: the party had supported the bill in Parliament, so why would it not attend the function that launches the tax?
The tax has had a long road to this hour, and parties on both sides of the hall have travelled it. This chapter is about how a country with that kind of politics managed to pass a reform of this size at all, and what it cost to do it.
At midnight the Goods and Services Tax comes into force. It replaces, at the centre, the central VAT, the service tax, the central sales tax and two kinds of additional customs duty, and in the states the VAT, sales tax, entry tax, luxury tax and entertainment tax. The rates are set at 0, 5, 12, 18 and 28 percent, and firms with turnover under two million rupees stay outside it. A government feature published on the day promised that an end to multiple taxation would make one market, and quoted estimates of a gain of up to two percentage points of GDP.
The word for what happens at midnight is not “passed”. The law passed a year earlier. What happens now is the first day of a bargain, and the bargain is older than anyone in the hall expects.
The bargain begins with the forum chapter 15 left in 2000: the Empowered Committee of State Finance Ministers, set up under the Vajpayee government and chaired by the Left Front’s finance minister in West Bengal. The body that would design the new tax was a standing forum of the states’ own finance ministers, and the states, not the Centre, owned the design.
State VAT was the first thing that body built. The budget speech of 2005-06 said all the states had agreed to introduce it from 1 April 2005, and named a national GST as the goal of the UPA government. It was a trial of the administration that would be needed later, and of the trust. The 2006 budget then proposed 1 April 2010 as the start date for the GST. The date came and went, because no one had settled what the states would be paid for what they gave up.
The Constitution (115th Amendment) Bill was introduced in the Lok Sabha on 22 March 2011. The Standing Committee on Finance, chaired by Yashwant Sinha, reported only on 7 August 2013, more than two years after the bill was introduced. The committee had examined the bill, and the bill still did not move, which tells us that a committee review alone did not carry the 2011 version. In the meantime the resistance was led by two states governed by the party that would later pass the tax: Madhya Pradesh and Gujarat. They spoke for state autonomy, and other states shared the worry. The UPA government accepted most of the committee’s changes and offered higher compensation, and still could not move the bill.
Nor was Parliament in a state to carry it. Disruptions had escalated from the Winter Session of 2010, over the demand for a joint committee on the 2G spectrum affair, and since then the Lok Sabha had worked 52 percent of its scheduled time. The Monsoon Session of 2012 scheduled 120 hours for the Lok Sabha and used 24. The 15th Lok Sabha finished with productive time of 61 percent, the worst in more than fifty years, against 87 percent for the 14th and 91 for the 13th. The 2013 Finance Bill and demands for grants worth Rs 16.6 lakh crore were passed without any discussion. This book tells the story of both camps with the same pen: the governments of the party that carried the tax in 2017 resisted the 2011 bill in two states, and the Congress, whose government introduced the bill in 2011, boycotted the launch in 2017.
What changed between the failure of 2011 and the success of 2016 was not the economics. The same tax was on the table. What changed was the offer. Several changes to the text answered the objections of the states, and each is a plain line of law.
The amendments of 2016 dropped a 1 percent tax on goods moving between states. They also provided for a dispute mechanism for the Centre and the states. They required Parliament to compensate the states by law for any loss of revenue from the GST, for a period of five years. And they set up a GST Council, with a vote that was weighted so that neither side could win alone. In the Constitution’s own words, every decision of the Council must be taken “by a majority of not less than three-fourths of the weighted votes of the members present and voting”. The Centre’s vote carries a weight of one-third of the total votes cast, and the votes of all the states taken together carry two-thirds.
Do the sum. Three-fourths of the votes is the threshold. The Centre holds one-third, which is less than three-fourths, so the Centre cannot pass anything alone. The states hold two-thirds, which is also less than three-fourths, so the states cannot either. But one-third is more than the one-quarter that blocks a motion, so the Centre can stop anything it opposes. The states can stop a motion only if enough of them vote together to hold more than a quarter of all votes. A motion passes only if the Centre agrees and a large majority of the states do too. Disputes between the Centre and the states, or among the states, go to a mechanism the Council itself must set up.
Put the structure in the terms of the 2011 objection. The fear was that states would lose fiscal powers without guarantees. The 2016 text answers that fear in three parts: a seat and a vote at the table, a collective veto, and a written duty on Parliament to make up a loss. Both sides took a limit on themselves so that each could hold the other to a promise.
The Rajya Sabha had stood in the way for a year, and back-channel talks led by Finance Minister Arun Jaitley with state governments cleared the standoff. The idea was taken up in 2000, scheduled in 2006, introduced in 2011, and in force in 2017: seventeen years after the first committee met.
Chain I242000 → 2017A states' committee built the tax; a weighted council passed it19472026Two other tables in 2016
The GST was not the only reform in these months that went through a body built for the purpose, and one of them is worth setting beside it.
The Insolvency and Bankruptcy Code, whose passage chapter 19 followed, went from introduction on 21 December 2015 through a joint committee to both houses by 11 May 2016, with an expert committee’s draft in hand. In 2019 Parliament amended it in a week and a half, tightening the time limits and setting a floor for what a supplier owed money would receive. The sequence was an expert draft, a joint committee and a short timetable, and the Lok Sabha, which had worked 24 of 120 scheduled hours in one session in 2012, passed a technical law in months once the process was built that way. Doing Business 2018, published on 31 October 2017, later put India’s rank for resolving insolvency up 33 places, from 136th to 103rd. The same listing is a reminder of how much was left undone: India stood 181st for dealing with construction permits and 164th for enforcing contracts. It is a series the World Bank later discontinued, after an investigation into data irregularities in the 2018 and 2020 editions, so the ranking is a pointer and not a measurement.
The second was a committee that took a power away from the person who had held it. Until 2016 the Reserve Bank’s governor set interest rates. On 27 June 2016 the government brought into force an amendment to the RBI Act creating a six-member Monetary Policy Committee: three members from the Bank, with the governor as chair, and three outside experts appointed on the recommendation of a search committee headed by the Cabinet Secretary. On 5 August it notified a consumer-inflation target of 4 percent, with a tolerance band of 2 to 6 percent, until 31 March 2021, finalised in consultation with Governor Raghuram Rajan. Rajan had told his staff on 18 June that he would return to academia when his term ended on 4 September, and said that the committee that would set policy was yet to be formed. The context was not calm. In May 2016 a ruling-party member of parliament, Subramanian Swamy, had written twice within a fortnight to the Prime Minister asking for Rajan’s removal, alleging among other things that high interest rates hurt small industry. The committee moved the decision from one person to six, with the government choosing three of them through a search committee. CPI inflation, which the World Bank puts at 10.0 percent in 2013, was 4.9 percent in 2016. In March 2026 the Centre kept the 4 percent target in place until 2031.
None of these three bargains was an act of unusual courage. Each was an act of architecture. The GST gave the states a vote and a written promise; the IBC and the MPC moved decisions to a body built for the purpose.
A glut across the border
China, in the same months, was dealing with the opposite kind of problem. India’s bargains were about getting something started. China’s were about getting something stopped.
On 5 March 2016 the Government Work Report put the 2015 result at 6.9 percent growth and made “supply-side structural reform” the organising idea. Its five tasks included cutting overcapacity and excess inventory. It reported that capacity cuts over the previous three years had removed more than 90 million tonnes of iron and steel and 230 million tonnes of cement. It said that overcapacity remained a serious problem in some industries, and it promised to deal with “zombie enterprises” through mergers, restructuring and bankruptcy, with 100 billion yuan of central support.
That is an admission of a glut, made in the country’s most formal annual document. It is also a different kind of reform from the GST. A closure is easier to announce than to carry out when the plant employs thousands of people and sits in a chain of local-government debt. In 2025 China exported a record 119 million tonnes of steel. The sources do not say how much of that steel came from capacity left open in 2016.
The contrast with India is the point of the juxtaposition, and it is a limited one. China reported its closures in an annual report. India wrote a voting rule into its constitution. Neither route is a clean success: the Chinese target sits beside a record export number, and the Indian rule, as the next sections show, left a fiscal bargain that wore away.
After the first day
The GST began in a hurry and with five rates, and the years since have been spent on the money. The Act provided compensation for five years. The five years were the whole of the insurance, and in June 2022 they ended with no successor. States’ revenue from the GST stayed below the roughly 3 percent of state GDP that the taxes it subsumed had raised.
2025In early September 2025 the Council agreed to move the GST to two main rates, 5 and 18 percent, with a separate 40 percent rate on a short list of “sin” goods; the new rates took effect on 22 September. It cut the tax on hundreds of goods, from food and insurance to air conditioners and school supplies. The BBC reported the move as meant to lift consumption and cushion the shock of the new US tariffs, and reported that states which depend on GST revenue worried about shortfalls. The Council could reform the tax quickly, within the weighted-vote rule. Whether it could also pay for the reform was a different question, because the guarantee that had made the original vote possible had run out.
2026The GST Council is the newer of two tables at which the Centre and the states divide money. The older is the Finance Commission, which the Constitution convenes every five years. The 16th, whose report was tabled on 1 February 2026 for the years 2026 to 2031, kept the states’ share of the divisible pool at 41 percent. Its formula weights income distance at 42.5, the 2011 population at 17.5 and demographic performance at 10. The commission’s award is fixed for five years at a time; the Council’s rates move whenever the weighted vote allows. The table worked. The money bargain had worn thin.
The table that was not built
The last part of this story is a failure, and it is the failure that shows what the success needed.
On 17 April 2026 the Lok Sabha defeated the Constitution (131st Amendment) Bill. Of the 528 members present and voting, 298 voted for it and 230 against, short of the two-thirds a constitutional amendment needs. The bill would have enlarged the Lok Sabha from 543 to 850 seats, removed the link between the next delimitation and the first census after 2026, and brought women’s reservation into effect using the 2011 census. Opposition parties argued that adding seats on the 2011 numbers would reduce the relative weight of the southern and north-eastern states, which had slowed their population growth. The government’s stated aims were to enlarge the House and bring women’s reservation into effect alongside delimitation.
The bill tied women’s reservation, which had passed Parliament in September 2023, to a contested change in the weight of the states. The bundle failed. This book will not say who was right about delimitation. It can say what the GST showed: a proposal that moves power between levels of government passed in 2016 when the level that stood to lose held a vote and a written promise. The 2026 bill moved the weight of the states without either, and failed. Two cases are not a law.
Over the same years the habits of Parliament changed. The share of bills referred to committees was 71 percent in the 15th Lok Sabha, 28 percent in the 16th and 16 percent in the 17th, lower than in any of the three before it. In the 17th, 58 percent of bills passed within two weeks of introduction, and four bills went to Joint Parliamentary Committees. In the 18th Lok Sabha there has been a rise in the number of bills sent to joint committees. The research for this book found no analysis showing that committee referral makes a law last; the link is observed, not proved. Two of the three reforms of 2016 that this chapter covers, the IBC and the GST amendment, were examined by committees before passage. The MPC was not.
The midnight session of 2017 looked like a triumph, and it was a real one. It was also the end of a long walk through committees, and the beginning of another about money. What the walk shows is simple. The tax passed when the states were given a table with a vote. The money bargain wore away when the promise ran out. The bundle failed when the states were asked to accept a change in their weight with neither.
Across the border, in the same years, a state with no such table was making a different kind of choice. In Beijing a ministry decided, by list, whose batteries a subsidised car could carry.