Tuesday, 14 January 1997. A quarter to eight in the morning, on the beach at Badagara, on the Kerala coast.
The beach market opens at five and is mostly finished by eight. The effective close of the market is at hand, and nobody on the sand will buy the late sardines. The price is zero. Eleven fishermen tip their catch back into the sea.
Within fifteen kilometres, at Chombala and Quilandi, sardines are fetching 9.9 and 9.8 rupees a kilo. Twenty-seven buyers will go home without any fish. An average boat carries 381 kilograms. The fuel to run to the better beach would cost about 205 rupees. Staying put at Badagara, by Jensen’s reckoning, costs a crew up to 3,400 rupees. Nobody on the water can know this. Fish here are sold by auction on the beach, and no fisherman can find out what the next beach is paying.
The economist Robert Jensen’s study of Kerala’s sardine fishery describes this morning. It is the right place to begin a chapter about India in ordinary years, because it shows what the state had not yet built and what it was about to allow others to build. The beach did not need a policy. It needed a number: what is the price over there, now?
Fifteen days later mobile service reached Kozhikode city, on 29 January 1997; it would take months for boats to adopt it, and Kannur followed on 6 July 1998 and Kasaragod city and Kanhangad on 21 May 2000. The towers reached twenty to twenty-five kilometres out to sea, which is where most sardine fishing is done. Nobody planned that. The service was not explicitly planned for fishermen; the towers served coastal cities, and the coast happened to be covered. Fishermen began to keep lists of dozens, then hundreds, of buyers’ numbers, to ring several before choosing where to land, and so in effect to hold an auction from the boat.
By 2001 more than six in ten of the fishing boats Jensen surveyed in three northern districts carried a phone, in a region where fewer than one person in twenty did. The spread of prices between beaches fell from 60 to 70 percent, measured as a coefficient of variation, to 15 percent or less. The 5 to 8 percent of each day’s catch that had gone unsold disappeared. Fishermen’s profits rose by about 8 percent, and consumer prices fell by about 4 percent.
Jensen’s result is a measured case of a cheap telephone doing real work for poor people without a factory. The rest of the chapter asks how India came to have the telephone at all. The answer is not a single brave decision. It is a run of ordinary decisions in the years between the great crisis of 1991 and the election of 2004, most of them made badly the first time, and the more interesting fact is that several of them were repaired.
1994New DelhiThe state was both the player and the referee. The Department of Telecommunications made the rules and also ran the incumbent network. In November 1994 it issued eight cellular licences for the four metros to the bidders who offered the most up front. From December 1995 it issued thirty-four more, to fourteen companies, across eighteen circles. Each licence carried a fixed fee. Payal Malik’s later reconstruction for LIRNEasia describes the early auctions as marked by overbidding, and puts what the licensees owed by 1999 at about $873 million. A fixed fee is a bet on how many customers you will have, paid in advance. Many of the bidders had bet too high. By the Supreme Court’s later account, the operators had consistently defaulted on their licence fees and asked for relief.
1999New DelhiIn 1999 India had 2.55 fixed telephone lines and 0.18 mobile phones for every hundred people. That is about one fixed line for every thirty-nine people, or one line of any kind for every thirty-seven. By September of that year India had 22.6 million working telephone connections, 570,000 public call offices and about 1.35 million cellular customers. Of its 6.07 lakh villages, 3.43 lakh had a public phone.
The government’s answer was the New Telecom Policy of 1999, approved on 26 March. On 22 July it offered the operators a way out: move from a fixed fee to a share of revenue, and pay 35 percent of the dues outstanding. The change took effect from 1 August. Twenty-one licensees migrated. They paid an interim 15 percent of gross revenue, and later 12, 10 and 8 percent of adjusted gross revenue, for the three categories of circle.
It was not uncontested. A public-interest petition went to the Delhi High Court, which on 10 August let the migration go ahead subject to the approval of the new Council of Ministers and the new Lok Sabha. The Cabinet considered the matter again on 21 October 1999, after the election, and approved it. The Comptroller and Auditor General, auditing the licences in 2016, called the package a bail-out for a slowing sector, to help service providers who were short of money. The growth that followed is documented too, though the sources do not show how much of it the package caused. This chapter keeps both descriptions.
The rescue came with a change in who judged disputes. In October 1999 a separate Department of Telecom Services took over the running of the network, which left the older department with policy and licensing. The regulator, the Telecom Regulatory Authority of India, was reconstituted by ordinance: the government now had to ask it for recommendations on the need for and timing of new operators and the terms of their licences. A new tribunal was set up to hear disputes between licensor and licensee, between operators, and between operators and groups of consumers. A further ordinance in January 2000 amended the regulator’s law again, after friction between the regulator and the department. On 1 October 2000 the operating wing was made a company, which became Bharat Sanchar Nigam Limited.
The chapter’s reading is that an operator facing a ministry that is also its competitor has a weaker case than one who can go to a tribunal; the sources show that the tribunal was created, not what it changed. That is the point of the sequence, more than any single rate.
2000New DelhiA third repair mechanism was out of sight, and took seventeen years to show. In 2000 the Empowered Committee of State Finance Ministers, chaired by West Bengal’s finance minister, who belonged to the Left Front, took up the sales-tax reforms, after uniform floor rates for state sales taxes were agreed in January.
Chain I242000 → 2017A forum the states own19472026The point of the committee was not the tax. It was that the states, not the Centre, owned the design, in a standing forum that outlasted any one government. All states agreed to introduce value-added tax from 1 April 2005, under a different government, which named a national goods and services tax as the goal. The 2006 Budget proposed 1 April 2010 for it. The date slipped. A constitutional bill in 2011 was resisted by Madhya Pradesh and Gujarat, both ruled by the BJP, on grounds of state autonomy, and stalled. The tax came into force at midnight on 1 July 2017, seventeen years after the committee’s work began in earnest. A Council Table tells how, and at what price to the states. Here the committee is enough: a body that survived changes of government and let the idea survive with it.
2002DelhiTuesday, 24 December 2002, Delhi. Prime Minister Atal Bihari Vajpayee inaugurates metro service between Tis Hazari and Shahdara, a line of 8.3 kilometres. He opens Kashmiri Gate station and rides the train to Seelampur, where he addresses a public meeting. His deputy, L. K. Advani, credits coordination between the Centre and the Delhi government. The minister for urban development says the line should carry about 200,000 passengers a day, and about 2.8 million a day on the full 63-kilometre first phase when it is done in 2005. PIB puts that phase at an estimated Rs 10,571 crore, excluding customs duty and works contract tax. Japan and South Korea have helped pay. The line opens to the public on Christmas Day.
Parliament had passed a special law for it in the same session, the Delhi Metro Railway (Operation and Maintenance) Act, replacing an ordinance. The Delhi Metro Rail Corporation was owned half by the Government of India and half by the Delhi government. E. Sreedharan led it, and it had a loan from the Japan Bank for International Cooperation. Phase I was finished in 2005, two years and nine months ahead of schedule and within budget.
ThePrint reports, on Sheila Dikshit’s recollection, that Vajpayee was told the project had been delayed for decades and answered “Please implement it immediately”, and adds that “it is said” he insisted on paying for his ride. What can be checked is smaller and enough: a corporation with a statute, a lender, a single leader and a date built a railway early in a city where that had not been expected.
The same state could do the opposite. In January 1999 Vajpayee had laid the foundation stone of the Golden Quadrilateral, the highway ring joining Delhi, Mumbai, Chennai and Kolkata, 5,846 kilometres in all. By 2003–04 it was paid for from a fuel cess of Rs 1.5 a litre on petrol and diesel, and from tolls. On 31 March 2004 only 2,612 kilometres of it were complete. The target had been December 2004. By the Economic Survey’s own account, the delays meant that the full economic impact of the highways was yet to come. It was declared complete in January 2012. A Metro line and a highway project were run by the same government within a few years of each other. One came early and one came late.
2003IndiaThe numbers moved slowly and then in one year. In May 2003 India introduced Calling Party Pays, under which the caller, not the person receiving the call, pays. An interconnection usage charge regime followed in 2003–04, and unified access licences in October 2003. On 2 July Reliance Infocomm, which had launched in December 2002, opened its Monsoon Hungama offer: a connection with a handset for Rs 501, including Rs 100 of talk time, on a three-year commitment, over a limited-mobility service. One account reports 1.5 million subscribers by that month.
Between 31 March 2003 and 31 March 2004 mobile connections rose from 13.00 million to 33.70 million, an increase of 159 percent, while fixed lines grew by less than 3 percent. In one year India added more mobile users than it had connected in all earlier years together. The mobile tariff was Rs 6.1 a minute in 1999–2000 and Rs 1.6 from March 2003. GSM tariffs, Rs 14.5 a minute in March 1998, reached Rs 0.77 by March 2004. In the quarter ending 31 December 2004 mobiles, at about 48 million, overtook 44.8 million fixed lines. India had passed the fixed line without ever having built it.
Keep the scale in view. In 2003 India’s teledensity was 6.6 for every hundred people and China’s was 42.3. By the World Bank’s count, mobile subscriptions per hundred people in 2004 were 4.6 in India and 25.7 in China. The Indian telephone story of this decade is a fast climb from a very low floor.
2003Gaya, BiharThursday, 27 November 2003, Gaya, Bihar. Satyendra Dubey turns thirty.
He is an officer of the Indian Engineering Service, on deputation to the National Highways Authority of India since July 2002 as project director at Koderma, on the Aurangabad–Barachatti stretch of the Golden Quadrilateral. For sixteen months he has been the man answerable for that section of road. He had made the contractor suspend three engineers over financial irregularities and rebuild six kilometres of substandard road. According to accounts of the project, he also wrote to the Prime Minister alleging that big contractors received inside information and passed work to firms without technical expertise.
He is coming back from a wedding in Varanasi. His car battery fails, and he takes a rickshaw. His driver finds him shot dead by the roadside in A. P. Colony.
The Central Bureau of Investigation registered a case on 14 December. On 1 February 2004 two suspects were found poisoned. In March 2010 a Patna court convicted three men. Whether he died for the road or for his wallet is disputed. The book takes no side, and the rest of this chapter does not depend on one.
What can be said is narrower. A young engineer who applied the rules strictly, and wrote to the top about it, was killed within sixteen months of arriving, and the Highways Authority changed its contract procedures afterwards. The sources do not show whether the change protected the next engineer. The rule at the end of this chapter is written for that gap.
2004Mumbai and DelhiThe growth of these years was uneven. By the official count GDP rose 6.1 percent in 1999–2000, 4.4 percent in 2000–01, 5.8 in 2001–02 and 4.0 in 2002–03, then 8.1 in 2003–04. Agriculture, the work of most Indians, grew 0.3 percent, fell 0.1 percent, grew 6.5, fell 5.2, and then grew 9.1. In 2004 India still had 56 percent of its workers in farming.
On 2 January 2004 the Sensex had closed at a record 6,026.59. In February the government dissolved the Lok Sabha eight months early and fought on a campaign it called India Shining. Polling ran in four rounds, from 20 April to 10 May. Counting began on Thursday, 13 May, and the Sensex surged on counting day and crashed shortly after. On Sunday the sixteenth, Sonia Gandhi was chosen to lead a coalition. On Monday the seventeenth the Sensex fell 565 points in a day, in a week when the Left said it would scrap the disinvestment ministry. Days later Gandhi recommended Manmohan Singh, the finance minister of 1991, and on Saturday the twenty-second he was sworn in as Prime Minister.
It is tempting to read the result as the country’s verdict on reform. The political scientist Ashutosh Varshney reads it differently. He argues that the 2004 defeat was driven more by regional alliances than by a verdict on reform. The new government kept the direction. Its Common Minimum Programme committed it to case-by-case privatisation and the retention of the navaratna public-sector companies. The pace changed, and the direction did not.
The decade had a shape, and this chapter’s claim is that the shape holds. Things that had a referee, a forum, a statute or a lender tended to be repaired. Things that had none tended to be left to a person’s courage. The cases that follow say how far it held.
The licensing rules were then changed in a way that raised the opposite problem. On 10 January 2008 the Department of Telecommunications decided to issue new licences on a first-come-first-served basis, with the cut-off date moved forward.
In November 2010 the Comptroller and Auditor General put the loss at Rs 1.76 lakh crore. It was a presumptive figure, an estimate of what auctions might have raised, not money that had been paid and lost. On 2 February 2012 the Supreme Court cancelled 122 licences and directed that they be auctioned. The pendulum had swung back to auctions, through audit and the courts and not through policy.
2017New DelhiOn 21 December 2017 a special court acquitted all the accused, saying the prosecution had failed to prove its case beyond reasonable doubt. That verdict concerned criminal guilt. It did not say whether the allocation had been wise. The Economic Survey of 2016–17 remarked on abundant caution in bureaucratic decision-making, favouring the status quo; it does not tie that to this case.
The chain is therefore a pendulum: auctions that bankrupted bidders in 1994, a rescue in 1999, an administrative allocation in 2008, auctions again in 2012, a court verdict in 2017 that left nobody sure what rule an honest official should follow.
At 31 March 2026 the Telecom Regulatory Authority of India counted 1,330.6 million subscribers, and a teledensity of 84.6 percent, excluding machine-to-machine connections. Private operators held 92.3 percent of subscribers, and the average revenue per gigabyte of wireless data was Rs 8.02, down from Rs 8.97 the year before.
Set that beside the baseline of 1999, when India had under three fixed or mobile lines for every hundred people. The 1999 migration did not cause the whole difference, and the sources do not apportion it. It is one input among several that the sector went on using.
In May 2004 the finance minister of 1991 had become Prime Minister. What his decade built on the factory floor, and on the ground beneath it, is the next part of this book.