The 2004 UPA shock verdict is usually explained in one sentence: rural India rejected the Bharatiya Janata Party and its India Shining campaign, and the Congress returned to office on the strength of that rejection. The sentence survives because it contains a fragment of truth arranged in the wrong shape. When counting closed on 13 May 2004, the Election Commission of India recorded the Congress as the largest party with 145 seats and the Bharatiya Janata Party second with 138, a result that almost every published forecast had inverted. Yet the Congress share of the national vote, 26.53 percent in the Election Commission of India statistical report on the 2004 general election, was lower than the 28.30 percent the party had won in defeat in 1999. A verdict in which the winning party loses vote share, gains 31 seats and assembles a government out of pre-poll allies, post-poll partners and outside backing is not a wave, and it cannot be read as one. This article defends a narrower and stronger claim. The 2004 verdict was sectoral, not generic. Its decisive seat movements were concentrated in identifiable state arenas where agrarian distress, state incumbency and alliance design converged, above all in Andhra Pradesh and Tamil Nadu, with Karnataka supplying the stress test that refines the argument. The post-2004 legislation on employment, information and rural health was the verdict’s policy translation, drafted in the Common Minimum Programme of May 2004 and carried through institutions built for that purpose. The verdict was sectoral, not generic; the post-2004 MGNREGA-RTI-NRHM legislation was the verdict’s specific policy translation.

The Shock of 13 May 2004: A Majority Assembled, Not a Wave Declared

The election that produced the shock was held in four phases between 20 April and 10 May 2004, after Atal Bihari Vajpayee’s government had advised an early dissolution of the Lok Sabha in February 2004, gambling on favourable state results from December 2003 and on a campaign built around growth, highways, telecommunications and the prime minister’s personal standing. The Bharatiya Janata Party and its allies entered the poll expecting a renewed mandate. The Congress entered it expecting, on the evidence of its own internal assessments reported in the press of March 2004, a fourth consecutive defeat. Counting day reversed both expectations within hours. By the evening of 13 May 2004 the Bharatiya Janata Party had fallen from 182 seats in 1999 to 138, and Vajpayee had conceded. The Congress, at 145 seats, had gained ground it had not predicted in a single internal projection that entered the public record.

The arithmetic of the majority that followed deserves slow reading, because later mythology compressed it into a single number and then argued about the number’s meaning. The brief for this series states the standard post-poll tally as 222 seats for the United Progressive Alliance against 185 for the National Democratic Alliance, built from Congress 145 plus allies 77, and Bharatiya Janata Party 138 plus allies 47. Other competent counts classify the same chamber differently: the Congressional Research Service report titled India’s 2004 National Elections counted 217 seats for the Congress-led alliance and 185 for the Bharatiya Janata Party-led alliance, and Andrew Wyatt’s analysis in Electoral Studies in 2006 counted 222 for Congress and its allies against 189 for the National Democratic Alliance. The discrepancies are not errors in the Election Commission record. The Commission counts parties and constituencies; alliances are political classifications imposed on that record, and the classification of parties such as the Samajwadi Party, which backed the government on some questions and contested the Congress fiercely in Uttar Pradesh, shifts the totals. What no classification alters is the structure: a Congress-led core in the range of 217 to 222 seats, an outside bloc of Left parties holding roughly sixty seats, and a defeated alliance whose own tally sat below 190 on every count. The government that took office on 22 May 2004 rested on that layered structure, and the layering is the first clue to the verdict’s character. It was assembled, piece by piece and state by state, rather than delivered as a single national instruction.

Yogendra Yadav, the Centre for the Study of Developing Societies analyst whose post-poll work anchored serious discussion of the result, named the problem in the title of his Economic and Political Weekly essay of December 2004, The Elusive Mandate of 2004. Yadav’s point, developed with Suhas Palshikar across the National Election Study of 2004 and the state chapters collected in Electoral Politics in Indian States, published by Oxford University Press in 2009, was methodological before it was political. A first-past-the-post system converts modest and uneven vote movements into dramatic seat movements, and a federation of state party systems converts national questions into twenty-eight separate arguments conducted in different languages of caste, region and incumbency. The analyst who averages those arguments into one national mood manufactures a mandate that no voter ever cast. Yadav’s caution cuts against the rural-backlash reading at its root, because that reading is precisely an averaged mood: it takes the undeniable fact of agrarian distress in the early 2000s, the undeniable failure of the India Shining slogan to describe that distress, and the undeniable defeat of the incumbent alliance, and welds the three into a single national act of rejection. The weld looks solid only from a distance. At constituency distance it comes apart, and the pieces have names, dates and margins.

Consider the national vote figures once more, this time as a pair. In the Election Commission of India statistical report on the 2004 general election, the Congress polled 103,408,949 votes for 26.53 percent and 145 seats; the Bharatiya Janata Party polled 86,371,561 votes for 22.16 percent and 138 seats. In 1999 the Congress had polled a larger share, 28.30 percent, and won 114 seats; the Bharatiya Janata Party had polled 23.75 percent and won 182. Both national parties therefore lost vote share between the two elections. Both lost it to different recipients in different states. The Bharatiya Janata Party’s losses were converted into seat losses where its 1999 allies had deserted or collapsed, above all in Andhra Pradesh and Tamil Nadu. The Congress’s smaller losses were converted into seat gains where its 2004 allies pooled votes efficiently, in those same two states and in Maharashtra, Haryana and Jharkhand. The seat map moved violently; the vote map moved modestly and in contradictory directions. Any reading of 2004 that begins with the seat map and reasons backward to a national state of mind will mistake the electoral system’s amplification for the electorate’s voice. The sectoral reading begins with the vote map and the state arenas, and asks a harder question: in which arenas, among which occupational and regional constituencies, did votes actually move, and what did the winners in those arenas promise the people who moved?

That question has a documentary answer, and the answer is what distinguishes this article’s claim from a mood. Between 22 May 2004 and September 2005 the government formed after the verdict adopted a written programme, built an advisory council to convert campaign-era civil society drafts into bills, and enacted statutes on rural employment and public information whose design features can be traced clause by clause to the arenas that had moved. Governments reveal their reading of a verdict in the legislation they prioritise under scarcity, and legislative time in a coalition’s first year is the scarcest resource in Indian politics. The Manmohan Singh government spent that resource on a rural employment guarantee, a transparency statute and a rural health mission, in that order of political effort. A government that believed it had been elected by a generic anti-incumbent mood would have spent the same year on visibility and patronage. A government that believed it had been elected by urban liberalisation sentiment, of the kind the India Shining campaign had addressed, would have spent it on the second generation of market reform. The expenditure pattern is evidence. It records what Sonia Gandhi, Pranab Mukherjee and the Congress negotiators who drafted with the Left in May 2004 believed the verdict had demanded, and the state-level record assembled in the sections that follow shows that their belief was anchored in the geography of the result rather than invented after it.

Two Readings of the 2004 UPA Shock Verdict

The rural-backlash reading deserves statement at its strongest before it is adjudicated, because it was held by serious people on evidence they could cite. In its strongest form, associated in the weeks after May 2004 with Congress spokespersons claiming a mandate for the party’s common man platform and with commentators such as those writing in Frontline and the Economic and Political Weekly through the summer of 2004, the reading runs as follows. The National Democratic Alliance government of 1998 to 2004 presided over an agrarian economy in visible distress: agricultural growth had slowed after 1997, the drought of 2002 had cut output sharply, rural credit had contracted as a share of bank lending, and farmer suicides in Andhra Pradesh, Karnataka, Maharashtra and Punjab had become a documented scandal rather than an anecdote. Against that record the ruling alliance campaigned on India Shining, a government-funded publicity exercise extended into a party slogan, which invited voters whose wells had failed and whose cotton had fetched less than its cost to celebrate a national mood of arrival. The invitation insulted the insulted, and the insult was punished. On this reading the verdict’s meaning is moral and national: the countryside, joined by the urban poor, rejected a governing style that mistook the stock market and the mobile telephone for the country.

Almost every factual premise of that reading is defensible, and the article does not contest them. The drought of 2002 was severe; the National Crime Records Bureau data on farmer suicides through 2003 and 2004 recorded thousands of deaths concentrated in the cotton and dryland belts; the India Shining campaign was real, expensive and tone-deaf in the districts where the distress was deepest. The reading fails not on its premises but on its inference, and the failure is visible in three places. First, the geography: a generic rural rejection of the Bharatiya Janata Party should have damaged the party across rural India with some uniformity, yet the party held or improved its position in the rural seats of Madhya Pradesh, Rajasthan, Chhattisgarh and Gujarat, winning 25 of 29 seats in Madhya Pradesh and 21 of 25 in Rajasthan on the Election Commission count. Second, the beneficiary: in Uttar Pradesh, the largest rural state, the seats the Bharatiya Janata Party lost were collected not by the Congress but by the Samajwadi Party and the Bahujan Samaj Party, parties whose appeal was caste-bloc and state-specific, and the Congress tally in the state fell from ten to nine. Third, the mechanism: the two states that destroyed the National Democratic Alliance’s majority, Andhra Pradesh and Tamil Nadu, were states where the alliance’s senior partners were not the Bharatiya Janata Party at all but the Telugu Desam Party and the All India Anna Dravida Munnetra Kazhagam, each carrying a state incumbency of its own, each facing a state-specific agrarian grievance, and each abandoned before the poll by allies whose departure rearranged the vote-pooling arithmetic. A reading that cannot explain why the backlash spared the Bharatiya Janata Party’s own heartland governments, enriched parties that were not part of the incoming government, and operated through the collapse of two regional incumbents is not an explanation of the 2004 verdict. It is a caption for it.

The sectoral reading replaces the caption with a mechanism. It holds that the decisive vote movements of 2004 occurred inside specific production sectors and the state arenas that organised them, above all the rain-fed agricultural sector of the Deccan and the south, where cultivators and agricultural labourers faced a compound crisis of water, credit, input prices and output prices in the years from 2000 to 2004. It holds that these movements were mediated everywhere by state incumbency, so that agrarian anger struck the party that held the state government, which in Andhra Pradesh and Karnataka meant two different national beneficiaries, and in Tamil Nadu meant the destruction of a state incumbent allied to Delhi’s incumbent. It holds that alliance design converted these sectoral movements into seats with unusual efficiency in the south, because the Congress in 2004 did what it had refused to do in 1999 and pooled its vote with the regional parties that commanded the sectoral constituencies. And it holds that the reading is confirmed by the policy sequence that followed, because the incoming government’s first statutes addressed the named sector with named instruments: wage employment on public works for the agricultural lean season, price and credit commitments for cultivators, a health mission aimed at rural districts, and an information statute whose earliest constituency was the rural works programme itself. Sectoral is a demanding word. It requires the analyst to name the sector, the arena, the movement and the instrument. The sections that follow name all four, state by state and statute by statute.

One clarification belongs at the threshold, because the sectoral reading is sometimes caricatured as a denial that anything else mattered in 2004. Elections are not single-cause events, and this article’s claim is about the verdict’s decisive margin and its policy content, not about the totality of motives in 390 million ballots. The complication named in the series brief, that anti-incumbency, post-Gujarat minority consolidation and anti-Hindutva mobilisation all moved votes in 2004, is addressed in full in a later section, and the answer given there is a distinction rather than a dismissal. The political-cultural forces decided much about which alliances could be built and which voters could be reached by them. The sectoral forces decided what the winning alliance’s voters in the decisive arenas were asking for, and the proof is that the sectoral content is what the alliance’s programme, drafted within a fortnight of the result, actually contained. A verdict can have a cultural frame and a sectoral substance. The error of the conventional reading is to mistake the frame for the substance because the frame is what television recorded.

The Evidence Standard: Seats, Votes and the Election Commission Record

The sectoral reading stands or falls on constituency evidence, and the evidence standard used throughout this article is the one the series brief mandates: the Election Commission of India statistical report on the 2004 general election for party and constituency figures, the CSDS National Election Study of 2004 for voter behaviour, and the programme and statute texts for the policy translation. The three sources answer different questions and must not be blended. The Election Commission record settles who won, where, and by what margin; it is silent on why. The National Election Study, directed by Yogendra Yadav and Suhas Palshikar with state teams whose chapters appeared in the 2009 Oxford University Press volume, samples why; its findings carry the uncertainty of survey evidence and are used here for direction and relative weight rather than for false precision. The programme and statutes settle what the incoming government did; they are the hardest evidence of how the verdict was read by the people who had to act on the reading.

Read under that standard, the five-state seat-shift table below is the article’s central exhibit. It sets out, for the five states with the largest seat movements between 1999 and 2004, the alliance pattern the Election Commission record shows in each year and the constituency pattern that gives each shift its meaning. Two comparability warnings attach to the table and are part of its evidence rather than footnotes to it. Bihar’s 1999 delegation of 54 seats was elected from the undivided state; Jharkhand was carved out in November 2000, so the 2004 comparison aggregates Bihar’s 40 seats with Jharkhand’s 14 to preserve the same territory. Uttar Pradesh’s 1999 delegation of 85 seats likewise included the five seats of the future Uttarakhand, and its 2004 delegation numbered 80. Alliance totals inside the table are classifications of Election Commission party results, stated with their components so that any reader can reclassify them and test whether the argument survives the reclassification. It does, because the argument rests on the party-level figures, which no classification touches.

The five-state seat-shift table:

State	Pattern in the Election Commission record, 1999	Pattern in the Election Commission record, 2004	Constituency pattern that gives the shift its meaning
Andhra Pradesh	Telugu Desam Party 29 and Bharatiya Janata Party 7 held 36 of 42 seats for the National Democratic Alliance; Congress held 5, and the All India Majlis-e-Ittehadul Muslimeen held 1	Congress 29, Telangana Rashtra Samithi 5 and the Majlis 1 made 35; the Communist Party of India and the Communist Party of India (Marxist) held 1 each; the Telugu Desam Party fell to 5, and the Bharatiya Janata Party won none	The reversal ran through the coastal districts, Rayalaseema and Telangana alike. Congress margins passed 130,000 votes at Visakhapatnam and 148,000 at Rajahmundry, and the Telugu Desam Party survived in the north coastal belt at Srikakulam, Bobbili and Anakapalli
Tamil Nadu	The Dravida Munnetra Kazhagam-led National Democratic Alliance held 26 of 39 seats; the All India Anna Dravida Munnetra Kazhagam and Congress combination held 13	The Democratic Progressive Alliance of the Dravida Munnetra Kazhagam, Congress, the Pattali Makkal Katchi, the Marumalarchi Dravida Munnetra Kazhagam and the Left held all 39; the All India Anna Dravida Munnetra Kazhagam and Bharatiya Janata Party combination won none	The Dravida Munnetra Kazhagam rose to 16 seats and Congress to 10, the Pattali Makkal Katchi and the Marumalarchi Dravida Munnetra Kazhagam crossed from the defeated 1999 alliance into the winning one, and all three Chennai seats returned Dravida Munnetra Kazhagam winners
Karnataka	Congress held 18 of 28 seats; the Bharatiya Janata Party held 7 and its ally the Janata Dal (United) held 3	The Bharatiya Janata Party rose to 18; Congress fell to 8; the Janata Dal (Secular) held 2, and the Janata Dal (United) won none	The incumbent Congress state government was punished in its own arena. The Bharatiya Janata Party, in opposition in the state since 1999, collected the anti-incumbent vote that agrarian distress generated, and the Janata Dal (Secular) held its Vokkaliga belt seats around Hassan and Mandya
Uttar Pradesh	In the undivided 85-seat state the Bharatiya Janata Party held 29, the Samajwadi Party 26, the Bahujan Samaj Party 14 and Congress 10, with smaller allies and an independent holding the remainder	In the 80-seat state the Samajwadi Party rose to 35, the Bahujan Samaj Party to 19, the Bharatiya Janata Party fell to 10 and Congress to 9; the Rashtriya Lok Dal held 3	Atal Bihari Vajpayee held Lucknow, Sonia Gandhi held Rae Bareli and Rahul Gandhi entered from Amethi, but the seats the Bharatiya Janata Party vacated were collected by the Samajwadi Party and the Bahujan Samaj Party, whose caste blocs, not the Congress organisation, owned the state’s rural constituencies
Bihar, with Jharkhand aggregated for 2004	In the undivided 54-seat state the Bharatiya Janata Party and Janata Dal (United) held 41 between them; the Rashtriya Janata Dal held 7 and Congress held 4	Across the two successor states the Rashtriya Janata Dal, Congress, Lok Janshakti Party, Jharkhand Mukti Morcha and Communist Party of India held 42 between them; the Bharatiya Janata Party and Janata Dal (United) fell to 12	Lalu Prasad Yadav held Chhapra for the Rashtriya Janata Dal, whose Bihar tally rose to 22; the Janata Dal (United) survived at Nalanda through Nitish Kumar and the Bharatiya Janata Party at Bhagalpur through Sushil Kumar Modi, pockets of survival inside a reversed state

Three features of the table discipline everything that follows. The first is asymmetry of scale: Andhra Pradesh and Tamil Nadu together moved 70 seats from one alliance column to the other between 1999 and 2004, a movement larger than the national majority margin of the incoming government, and in both states the movement ran against a regional incumbent rather than against the Bharatiya Janata Party as a direct contestant. The second is the Karnataka inversion: the state with the same drought, the same agrarian distress and the same southern geography moved in the opposite partisan direction, because the incumbent being punished there was the Congress itself, and any honest version of the sectoral reading must explain Karnataka rather than hide it. The third is the northern pattern: Uttar Pradesh and Bihar moved massively, but through caste-bloc parties and state incumbency contests whose content was Mandal-era social justice politics applied to agrarian constituencies, not through the Congress at all. The sectoral claim survives these features because it was never a claim that one party benefited everywhere. It is a claim about where the decisive movements occurred, what occupational world the moving voters inhabited, and what the eventual government understood itself to have been asked for.

Andhra Pradesh: The Fall of the Chief Executive Model

Andhra Pradesh is the cleanest case in the 2004 record, and K.C. Suri’s title for the state chapter in the CSDS volume of 2009 states the case in a phrase: Andhra Pradesh: Fall of the CEO in the Arena of Democracy. N. Chandrababu Naidu, chief minister since 1995, had built the most internationally celebrated state government in India, the government of Vision 2020, of Hyderabad’s technology parks, of power sector restructuring under World Bank guidance, and of a chief minister who presented himself, and was presented by the global business press between 1999 and 2003, as a chief executive rather than a politician. The Telugu Desam Party and Bharatiya Janata Party alliance he led won 36 of the state’s 42 Lok Sabha seats in 1999, and his government was returned in the simultaneous state election of that year with its reform credentials intact. By the election of April and May 2004 the same alliance held five seats, the Bharatiya Janata Party held none, and the Congress-led combination held thirty-five on the brief’s classification, thirty-seven if the two Left seats are added to the Congress column. No state in India moved further, and the movement’s content is documented in unusual detail because the grievance had been organised before it was counted.

The grievance was water, power and debt, in that order of daily experience. The monsoons of 2002 and 2003 failed across the state’s dryland districts; Anantapur, Kadapa, Mahbubnagar and the Telangana uplands entered the election after successive seasons of borewell failure, and the National Crime Records Bureau figures for 2003 and 2004 placed Andhra Pradesh at or near the top of the national table of farmer suicides, a ranking the state government disputed in method and could not dispute in political effect. Power sector reform, the Naidu government’s signature achievement in the eyes of its external admirers, had raised agricultural tariffs and tightened metering in a countryside where the pump set was the difference between a crop and a debt spiral. Institutional credit had thinned as cooperative banks weakened, and the gap was filled by private moneylenders at rates that converted one failed season into a land sale. Y.S. Rajasekhara Reddy, the Congress leader in the state, spent April to June 2003 walking the affected districts on a padayatra of some 1,500 kilometres from Chevella to Ichchapuram, an exercise whose entire method was to collect the grievance village by village and attach it to a short list of promises: free power for agriculture, loan relief, and irrigation investment on a scale the incumbent had deferred. The list matters for this article’s argument because it is the sectoral demand written as a campaign document a full year before the national verdict, in the state that would supply the verdict’s largest single bloc of moved seats.

What did Y.S. Rajasekhara Reddy promise farmers on electricity in 2004?

Y.S. Rajasekhara Reddy promised free electricity for agricultural pump sets, reversing the tariff increases of the Naidu government’s power reforms. The Congress manifesto for Andhra Pradesh in 2004 paired the promise with farm loan relief and expanded irrigation, and Reddy signed the free power order on taking office as chief minister in May 2004.

The alliance design converted the grievance into seats with a completeness that the 1999 design had achieved for the other side. In 1999 the Congress had contested all 42 seats alone and won five on a vote share, recorded in the Election Commission of India statistical report on the 1999 general election, of 42.79 percent, a plurality converted into almost nothing by the opposing alliance’s pooling. In 2004 the Congress contested in harness with the Telangana Rashtra Samithi of K. Chandrashekar Rao, whose demand for a separate Telangana state had broken from the Telugu Desam Party in 2001 and whose five seats in the 2004 count were all in the Telangana region, and with the Left parties in a seat adjustment that removed the three-cornered contests of 1999. The Congress vote share in the state actually fell slightly, to roughly 41 percent on the Election Commission count, and its seat tally rose from five to twenty-nine. No single exhibit in the 2004 record demonstrates the article’s methodological point more cleanly: the sectoral grievance was the fuel, but the alliance was the engine, and the conventional reading, which speaks of backlash as though votes converted themselves, has no vocabulary for the engine at all. The simultaneous state election completed the picture. The Congress won 185 of 294 assembly seats, the Telugu Desam Party fell to 47, and Naidu resigned on 11 May 2004, two days before the national counting finished, because the state verdict that carried the Lok Sabha verdict inside it was already beyond argument.

The constituency pattern inside the state confirms the sectoral content rather than a diffuse anti-incumbency. The margins in the Election Commission record run largest in the agrarian constituencies of the coast and the interior: Nedurumalli Janardhana Reddy’s margin at Visakhapatnam exceeded 130,000 votes, Vundavalli Aruna Kumar’s at Rajahmundry approached 149,000 against a Bharatiya Janata Party runner-up, and Congress winners took seats across Rayalaseema, the home terrain of both Naidu’s party organisation and Reddy’s own factional base. Asaduddin Owaisi held Hyderabad for the All India Majlis-e-Ittehadul Muslimeen, the single seat the party had also held in 1999, a continuity that marks the urban Muslim constituency’s separate logic inside the state pattern. The Telugu Desam Party’s five survivals clustered where its local machines were oldest. Suri’s chapter in the 2009 CSDS volume, drawing on the National Election Study state sample, records the class gradient that the sectoral reading predicts: the Congress-led alliance ran strongest among agricultural labourers and the very poor, with the state sample recording roughly half of very poor respondents and a majority of women respondents behind the alliance, and weakest among the peasant proprietor groups where the Telugu Desam Party retained a residual hold. A generic backlash does not produce a class gradient of that shape. A sectoral verdict does, because the sector in distress was not agriculture as an abstraction but the bottom of agriculture: the labourer without land, the tenant without title, the smallholder without institutional credit.

Tamil Nadu: A Complete Alliance Reversal in Thirty-Nine Seats

Tamil Nadu supplied the second destruction, and its completeness still startles: thirty-nine seats contested, thirty-nine seats won by the Democratic Progressive Alliance, zero by the combination of the All India Anna Dravida Munnetra Kazhagam and the Bharatiya Janata Party that had expected, on the reporting of April 2004, to carry the state. The brief for this series describes the winning core as the Dravida Munnetra Kazhagam, Congress and Pattali Makkal Katchi alliance taking 35 of 39 seats; the fuller Election Commission classification adds the Marumalarchi Dravida Munnetra Kazhagam’s four seats to the same alliance and takes the total to the full thirty-nine. In 1999 the state’s alignment had been the mirror image. M. Karunanidhi’s Dravida Munnetra Kazhagam had anchored the National Democratic Alliance in the state, and that alliance’s 26 seats had included the Pattali Makkal Katchi and the Marumalarchi Dravida Munnetra Kazhagam alongside the Bharatiya Janata Party. The 2004 verdict was therefore not a contest between a standing government bloc and a standing opposition bloc. It was a re-sorting, in which two caste-anchored regional parties carried their voters across the alliance line and the state’s entire delegation changed hands around them.

The incumbent being punished in Tamil Nadu was J. Jayalalithaa’s state government, in office since May 2001, and the grievances against it had accumulated along lines that the sectoral reading predicts and the national reading cannot see. The state had passed through severe drought in 2002 and 2003; the Cauvery dispute with Karnataka had embittered the delta districts in the season of 2002 to 2003, when Karnataka’s Congress government under S.M. Krishna restricted releases and Tamil Nadu’s paddy belt watched a kuruvai crop fail; and the Jayalalithaa government’s fiscal consolidation had withdrawn or tightened benefits that Tamil Nadu’s electorate had come to treat as entitlements of the Dravidian welfare settlement, including the free electricity supply to farmers that her government curtailed on taking office and the subsidised rice and festival provisions that opposition speakers itemised through the campaign of early 2004. Layered over the material grievance was a style grievance with material consequences: the anti-conversion legislation of 2002, the arrests and confrontations with government employees during the strike of July 2003, when the state dismissed large numbers of striking staff before the courts intervened, and a reputation for inaccessible government that Karunanidhi’s campaign converted into a simple promise of restoration. Restoration of what, is the analytical question, and the Dravida Munnetra Kazhagam manifesto of 2004 answers it in sectoral terms: farm power, water, welfare entitlements and the defence of the state’s claims against the Centre and against Karnataka.

The alliance mechanics repay attention because they show the sectoral vote being pooled with a precision that 1999 had lacked on the Congress side. The Dravida Munnetra Kazhagam took sixteen seats, the Congress ten, the Pattali Makkal Katchi five in the state with a sixth in Pondicherry, the Marumalarchi Dravida Munnetra Kazhagam four, and the two Communist parties two each. Each component owned a recognisable constituency: S. Ramadoss’s Pattali Makkal Katchi commanded the Vanniyar belt of the northern districts, Vaiko’s Marumalarchi Dravida Munnetra Kazhagam held pockets of the south, the Left parties held unionised and delta constituencies, and the Congress contributed a statewide residue plus the candidacies of senior national figures, among them P. Chidambaram, returned from Sivaganga, and Mani Shankar Aiyar, returned from Mayiladuthurai. Dayanidhi Maran’s victory in Chennai Central and T.R. Baalu’s in Chennai South showed the alliance sweeping the metropolitan seats as thoroughly as the delta, a fact sometimes deployed against the sectoral reading, as though urban victories dissolved the rural argument. The deployment confuses the location of a constituency with the composition of its electorate and the content of its campaign. Chennai’s seats were won on the state government’s record and the alliance’s welfare promises in a state where the boundary between the metropolitan voter and the agrarian hinterland runs through every extended household; the National Election Study state sample recorded the alliance ahead across locality types, which is what a state-incumbency verdict pooled by a complete alliance looks like. The sectoral claim for Tamil Nadu is not that only farmers voted the alliance in. It is that the alliance’s offer, and the grievance it organised, were built from the agrarian and welfare materials of the state’s political economy, and that the Bharatiya Janata Party, contesting six seats as the junior partner of an unpopular state incumbent, was destroyed as collateral in a state argument it barely featured in. That distinction, between being rejected and being collateral, is exactly what the generic reading erases.

The Tamil Nadu result also disciplines the national story through its sequel inside the incoming government. The Dravida Munnetra Kazhagam entered the Union cabinet with portfolios that touched the sectoral economy directly, and the state’s weight in the coalition’s arithmetic, thirty-nine seats delivered as a single bloc by Karunanidhi’s alliance management, made Tamil Nadu’s welfare expectations a standing term of the coalition’s internal bargain rather than a regional footnote to it. When the Common Minimum Programme was drafted in the second half of May 2004, its commitments on employment, food security and social protection were written by negotiators who knew precisely which state delegations had made the government and what those delegations’ electorates had been promised at home. The sectoral reading follows that chain from the delta districts to the drafting table. The generic reading stops at the slogan.

Karnataka: The Stress Test That Strengthens the Sectoral Reading

Karnataka is the state the conventional accounts of 2004 mention least, and the state this article treats as its most instructive, because Karnataka breaks every simple version of the story including the simple version of this article’s own thesis. The Election Commission record is stark. In 1999 the Congress held eighteen of the state’s twenty-eight Lok Sabha seats and governed the state under S.M. Krishna; the Bharatiya Janata Party held seven and its Janata Dal (United) ally three. In 2004 the Bharatiya Janata Party won eighteen seats, the Congress fell to eight, the Janata Dal (Secular) of H.D. Deve Gowda held two, and the Janata Dal (United) won none. Sandeep Shastri and Harish Ramaswamy titled their chapter in the 2009 CSDS volume Karnataka: Simultaneous Polls, Different Results, because the simultaneous assembly election produced a fractured verdict in which the Bharatiya Janata Party emerged largest with 79 seats, the Congress fell to 64 and the Janata Dal (Secular) took 59, and the two verdicts, delivered by the same voters on the same day, diverged in instructive ways. A series brief that describes Karnataka as a Congress success misreads the record; the record shows a Congress collapse. The sectoral reading is strengthened, not weakened, by stating the collapse plainly and explaining it, because the explanation is the thesis in its precise form.

The explanation begins with incumbency and sector, the two variables the thesis actually contains. Karnataka entered 2004 after four consecutive years of drought, recorded in state government reporting from 2000 onward and seared into the state’s politics by farmer suicides in the northern districts and by the Cauvery confrontation of 2002, in which Krishna’s government was seen by the state’s farmers to have yielded water to Tamil Nadu under judicial and political pressure. Krishna’s government had a genuine record of urban administration in Bengaluru and a national reputation as a modernising chief minister, and it carried, fairly or not, the charge of governing for the capital’s technology economy while the northern countryside dried. The charge is the Andhra charge with the parties reversed: a celebrated chief executive model, a globalised enclave, an agrarian interior in distress. The distress therefore punished the incumbent, and the incumbent in Karnataka was the Congress. The beneficiary was the party of opposition in the state, the Bharatiya Janata Party, whose organisation had been built across the state’s Lingayat and coastal belts during the 1990s and whose tally rose from seven seats to eighteen on a vote share that the Election Commission recorded rising from roughly 27 percent in 1999 to roughly 35 percent in 2004. The Janata Dal (Secular), the third force, held the Vokkaliga belt seats of the south and emerged from the assembly verdict as the Congress’s coalition partner in the state government that followed, under N. Dharam Singh from May 2004.

What Karnataka refutes is the proposition that agrarian distress in 2004 carried a fixed partisan charge against the Bharatiya Janata Party or for the Congress. What it confirms is the sectoral mechanism stripped of party labels: distress moved against incumbents in the arenas where it was lived, through the opposition organisations locally available to receive it, and the national consequences were the sum of those local transactions rather than the expression of a national preference. This is why the article insists on the word sectoral rather than rural. A rural reading predicts that the countryside spoke with one voice and asks only how loudly. A sectoral reading predicts that cultivators and labourers in distress punished the governments answerable for water, power, credit and prices in their own states, and it is corroborated when the punishment falls on a Congress government in Bengaluru and a Telugu Desam government in Hyderabad with the same force. The policy translation that followed in Delhi then requires a further step, and the step is documented rather than assumed: the Congress leadership in Delhi read the national sum of these state transactions, including the loss of its own Karnataka government, as a verdict on the agrarian settlement of the preceding decade, and wrote that reading into the Common Minimum Programme. Politicians are permitted to learn from their own defeats as well as from their victories, and the Karnataka defeat, which cost the Congress its most admired state government, may have taught the lesson more sharply than the Andhra victory did.

Uttar Pradesh and Bihar: The Northern Shifts and Their Different Logics

The northern evidence performs a double service in this argument. It confirms that 2004 was a year of massive seat movement against the National Democratic Alliance’s 1999 position, and it refuses, on inspection, to let that movement be described as a Congress revival or as a single rural mood. In Uttar Pradesh the Election Commission record for 2004 shows the Samajwadi Party at thirty-five seats, the Bahujan Samaj Party at nineteen, the Bharatiya Janata Party at ten and the Congress at nine, in an eighty-seat state whose undivided predecessor had returned twenty-nine Bharatiya Janata Party members in 1999. Mulayam Singh Yadav’s Samajwadi Party, which had taken the state government in August 2003 after the collapse of Mayawati’s coalition with the Bharatiya Janata Party, converted incumbency, a disciplined Yadav and Muslim constituency base, and a network of local strongholds into the state’s dominant tally. Mayawati’s Bahujan Samaj Party converted its Dalit base and a widened candidate strategy into nineteen seats on a vote share of roughly a quarter of the state’s ballots, second only to the Samajwadi Party. The Congress, by contrast, stagnated in the state that had once been its citadel: Sonia Gandhi’s margin in Rae Bareli and Rahul Gandhi’s first victory in Amethi were personal and dynastic holdings inside a state organisation that won nothing else beyond seven scattered seats, and the party’s state vote share of roughly twelve percent was its weakest in the large states.

The analytical weight of Uttar Pradesh lies in what the state’s result did to the national government’s composition. The Samajwadi Party and the Bahujan Samaj Party together held fifty-four seats, a bloc larger than the Left’s, and neither party entered the United Progressive Alliance. Mulayam Singh Yadav’s relations with the Congress were adversarial in the state both parties contested, and his party’s backing for the government in Delhi was episodic and priced vote by vote. A verdict that had truly been a national instruction to install a Congress-led redistribution programme would presumably have strengthened the Congress in the largest state. Instead the largest state strengthened two parties whose programme was the Mandal-era settlement itself: representation and state resources for defined caste constituencies, pursued through control of the state government and its employment, police and land machinery. The connection to the sectoral thesis is real but must be stated exactly. The constituencies that moved in Uttar Pradesh were overwhelmingly agrarian, and the parties that received them organised agrarian caste blocs; the content of the demand, however, ran through caste representation and state-level control rather than through the price, water and credit grievances that organised Andhra Pradesh. The sectoral reading accommodates this by treating sector and arena together: in the Andhra arena the operative unit was the distressed cultivating economy, in the Uttar Pradesh arena it was the caste bloc inside that same economy. The generic reading, which needs a single national demand, accommodates neither.

Bihar’s reversal was larger in proportion and older in lineage. Aggregating the successor states for comparability, the Election Commission record moves from forty-one of fifty-four seats for the Bharatiya Janata Party and Janata Dal (United) in 1999 to twelve in 2004, and from eleven seats for the Rashtriya Janata Dal and Congress in 1999 to forty-two for the wider Congress-aligned combination in 2004, built from the Rashtriya Janata Dal’s twenty-two Bihar seats and two in Jharkhand, the Congress’s three in Bihar and six in Jharkhand, the Lok Janshakti Party’s four and the Jharkhand Mukti Morcha’s four, with the Communist Party of India holding Hazaribagh. Lalu Prasad Yadav, whose party governed Bihar throughout the period and whose own Chhapra victory anchored the tally, had built the Rashtriya Janata Dal’s constituency on the same compound that the sectoral reading identifies elsewhere: an agrarian electorate organised by caste, addressed through social justice claims on the state, and hostile to the alliance that represented the state’s older dominant groups in Delhi. Ram Vilas Paswan’s Lok Janshakti Party brought a Dalit constituency into the winning combination that it had carried to the National Democratic Alliance in 1999, a crossing that mirrors the Pattali Makkal Katchi’s crossing in Tamil Nadu and underlines how much of 2004 was decided by the movement of organised blocs between alliances rather than by unorganised opinion drifting in a national wind. Nitish Kumar’s survival at Nalanda and Sushil Kumar Modi’s at Bhagalpur mark the pockets where the Janata Dal (United) and Bharatiya Janata Party machines held, and their isolation in the tally measures the completeness of the reversal around them.

For the Delhi government’s policy translation, Bihar and Uttar Pradesh posed a problem that the southern victories did not. The programme drafted in May 2004 had to speak to agrarian constituencies that the Congress did not itself represent in the two largest states, through instruments that only a Union government commands: employment law, food distribution, rural credit targets, health missions and the financing of state schemes. The Common Minimum Programme’s emphasis on legal entitlements rather than on discretionary schemes is partly explained by that problem. An entitlement travels across state boundaries and party organisations; a scheme is captured by the state machinery that administers it. The negotiators of May 2004, among them Pranab Mukherjee for the Congress and Sitaram Yechury and Prakash Karat for the Communist Party of India (Marxist), were drafting for a country in which the government’s own party was weak exactly where rural India was most numerous, and the universalist, rights-bearing shape of what they drafted carries that weakness inside it as a design principle.

The Cultural Countercase: Gujarat, Minorities and the Terms of Alliance

The strongest objection to the sectoral reading is not statistical but moral and political, and it must be met on its own ground. The objection holds that 2004 was decided substantially by the cultural politics of the preceding years: by revulsion at the Gujarat violence of February and March 2002, in which more than a thousand people, most of them Muslims, were killed after the burning of a train coach at Godhra on 27 February 2002 killed fifty-nine Hindu pilgrims and activists; by the consolidation of Muslim voters behind whichever alliance could defeat the Bharatiya Janata Party in their constituencies; and by a broader anti-Hindutva mobilisation conducted by the Left parties, by civil society campaigns after Gujarat, and by the Congress’s own campaign rhetoric under Sonia Gandhi, who had attacked the Vajpayee government’s record on communal harmony throughout 2003 and into the poll. The Communist Party of India (Marxist), in its Central Committee review of the elections adopted between 30 July and 1 August 2004, stated its three objectives in exactly this register: defeat the Bharatiya Janata Party and its allies, install a secular government, and increase the Left’s strength. Those objectives were achieved, and the review’s language is evidence that the winning side’s own organisations understood the contest substantially in cultural terms.

How many Gujarat seats did the Bharatiya Janata Party and the Congress win in 2004?

The Bharatiya Janata Party won fourteen of Gujarat’s twenty-six seats in 2004 and the Congress won twelve, on the Election Commission of India count. The Congress gain of six seats on its 1999 tally was the party’s best Gujarat performance in a generation, and the Bharatiya Janata Party retained the state it governed.

The Gujarat figures inside the objection deserve the same care as the objection itself. The Congress rise from six seats in 1999 to twelve in 2004 is consistent with a Muslim consolidation joined by other constituencies in the state’s closely divided politics, and constituency studies of the period record heavy anti-Bharatiya Janata Party voting by Muslim electorates across north and central India. Yet the same record shows the Bharatiya Janata Party retaining Gujarat, the state where the cultural argument should have bitten hardest, and retaining it under Narendra Modi, the chief minister whose government’s conduct in 2002 was the objection’s central exhibit. The cultural forces were real, and their seat yield was bounded by the same state arenas that bounded everything else in 2004. Muslim consolidation could move seats where a viable alternative alliance existed to receive the consolidated vote, as in parts of Uttar Pradesh, Bihar and West Bengal; it could not by itself manufacture an alternative in Gujarat, and it did not determine the content of the programme the winners enacted, which contained no statute addressed to the cultural grievance and five major commitments addressed to the agrarian and welfare economy.

The distinction this article draws is therefore between the conditions of the victory and the content of the verdict. The cultural forces were conditions: they decided that the Dravida Munnetra Kazhagam could leave the National Democratic Alliance without paying a secular penalty, that the Lok Janshakti Party could cross to the Congress side, that the Left’s sixty seats would be available to sustain a Congress government rather than to bargain with a chastened Vajpayee, and that Muslim voters in mixed constituencies would coordinate behind the strongest local alternative. Anti-incumbency operated as a second condition, general in direction and local in incidence, punishing state governments from Hyderabad to Bengaluru to Chennai and the Union government in Delhi on a schedule set by each arena’s own grievances. Conditions decide who can win. Content decides what the winner is asked to do, and content is read from what the moving constituencies demanded in their campaigns and what the winners then legislated under scarcity. In Andhra Pradesh the demand was farm power, water and credit. In the Common Minimum Programme the answer was employment, agriculture and social protection. The chain from demand to answer is the sectoral reading, and no part of the cultural countercase breaks a link in it. The countercase explains why the door was open. It does not explain what walked through.

From Counting to Claiming: Sonia Gandhi, Manmohan Singh and 22 May 2004

Between the counting of 13 May and the swearing-in of 22 May 2004, the verdict acquired its human form, and the choices of those nine days shaped how the sectoral demand would be governed. The Congress Parliamentary Party elected Sonia Gandhi as its leader on 15 May 2004, and allies and partners delivered letters of backing that placed the majority within her reach. On 18 May 2004 she announced to a stunned Congress Parliamentary Party meeting that she would not take the prime ministership, saying that she had listened to her inner voice and that the office had never been her aim. The Bharatiya Janata Party figures who had campaigned against her Italian birth, among them Sushma Swaraj, who had threatened in April 2004 to shave her head and sleep on the ground if Gandhi became prime minister, rendered their objection at its strongest as a claim about the symbolic ownership of the republic’s highest office; Gandhi’s defenders, and President A.P.J. Abdul Kalam’s later account in his 2012 memoir Turning Points, record that the presidency was ready to swear her in and that the refusal was hers. The motives remain disputed between those accounts, and the dispute is marked here as disputed. The institutional consequence is not disputed. By declining the office and retaining the alliance’s leadership, Gandhi separated the coalition’s political authority from the government’s executive head, and installed in the executive office the one Congress figure whose biography embodied the programme the coalition was about to write.

When did President A.P.J. Abdul Kalam appoint Manmohan Singh as Prime Minister?

President A.P.J. Abdul Kalam appointed Manmohan Singh as Prime Minister on 19 May 2004, after Sonia Gandhi presented letters of backing for a Congress-led government and named Singh as its head. Singh and his council of ministers were sworn in at Rashtrapati Bhavan on 22 May 2004.

Manmohan Singh’s qualifications for the role, as the Congress presented them in May 2004 and as his record supported, were those of the finance minister of 1991 to 1996 who had liberalised the economy under P.V. Narasimha Rao, a Rajya Sabha member from Assam with no personal electoral machine, and an administrator whose reputation for personal integrity survived every partisan attack of the preceding decade. The appointment is sometimes narrated as an accident of Gandhi’s refusal. Read against the programme negotiations running in parallel, it looks more like a matching of instrument to task. The government being assembled needed a prime minister who could negotiate with the Left without being captured by the argument, reassure the markets that had reacted nervously to the verdict and to Left statements in the week after 13 May 2004, and preside over a redistribution programme whose credibility depended on fiscal competence being visible at its head. Singh satisfied the three requirements simultaneously, and the economic team assembled around him, P. Chidambaram at Finance, Montek Singh Ahluwalia returning as Deputy Chairman of the Planning Commission in June 2004, and Pranab Mukherjee at Defence with the coalition’s political negotiations in his portfolio, signalled the same matching. The contrast with the outgoing government’s predicament is the contrast the 1999 analysis identified in the BJP’s strategic miscalculation that the 2004 verdict exposed: the Vajpayee government had won in 1999 by broadening its alliance faster than its programme, and by 2004 its partners had priced its programme state by state and left. The incoming government inverted the sequence. It wrote the programme first, in the open, with the partners’ signatures on it, and built the ministry around the text.

The outside backing that made the ministry possible had a pedigree the participants themselves cited. The Left parties, holding their largest Lok Sabha contingent in the republic’s history, chose to sustain the government without entering it, replicating at a larger scale the arrangement examined in the outside-support coalition model the UPA inherited. The inheritance was practical rather than ceremonial. The United Front years had taught the Congress that outside backing without a written programme and a consultative forum ends in a withdrawal letter; the 2004 arrangement therefore began with a negotiated text and a coordination committee, and the Left’s price for its votes was written into that text in the presence of witnesses. The wider frame within which the whole transaction sat is the one the pillar hub states as the coalition-era renegotiation framework: after 1989, Indian governments are standing renegotiations among parties that keep their separate electorates, and the quality of a coalition is measured by whether its renegotiations are conducted through institutions built for the purpose or through crises improvised under pressure. The events of May 2004 were the most deliberate piece of institution-building in that renegotiation history, and their deliberateness is a further reason to read the verdict through its translation. A government improvised in panic does not draft its programme before it drafts its cabinet list.

The Common Minimum Programme: The Verdict Translated into a Governing Text

The National Common Minimum Programme, released on 27 May 2004, five days after the government took office, is the single most probative document in this article’s argument, because it is the point at which interpretation hardened into commitment. The text was negotiated between the Congress and its allies and the Left parties in the days after 13 May 2004, with Pranab Mukherjee central to the Congress side of the drafting and the Left’s conditions, on employment, on disinvestment restraint, on agriculture and on social spending, incorporated as clauses rather than acknowledged as pressures. Its employment clause committed the government to enact a National Employment Guarantee Act providing a legal guarantee of at least one hundred days of employment a year, on asset-creating public works at minimum wages, to at least one able-bodied person in every rural household and in urban poor and lower middle-class households, with a food-for-work programme in the interim. Its agriculture clauses committed the government to enhanced public investment, expanded institutional credit and a flow of technology to farming and its allied activities. Its social clauses set targets that entered the period’s vocabulary: six percent of gross domestic product for education, a rise in health spending toward two to three percent, universalisation of the Integrated Child Development Services, a strengthened public distribution system, and social security legislation for unorganised workers. P. Chidambaram, presenting the Union budget on 8 July 2004, called the programme the government’s guiding light and organised the budget’s seven stated economic objectives around it, an act of subordination of the finance ministry to a coalition text that had no precedent in the Rao or Vajpayee years.

The constitutional resonance of the document was conscious, and its drafters worked a seam in the founding text that this series has followed from the start. The employment guarantee took the Directive Principle on the right to work, a non-justiciable instruction to the state since 1950, and converted a portion of it into a statutory entitlement enforceable by a rural household against its own panchayat. That conversion is the directive-rights tension that MGNREGA renegotiated rendered as legislation, and behind it stands the longer arc of the rights-development founding tension: the founders’ decision to place material citizenship in the non-enforceable part of the Constitution, on the argument that resources must precede rights, met in 2004 a coalition whose electoral survival depended on reversing the sequence for one named sector. The agricultural clauses reached back along a different line, to the rural-agricultural political crisis precedent, when food scarcity last reorganised national politics and taught the Congress that the countryside’s grievances, left unpriced, return as verdicts. The drafters of May 2004 had all lived through the verdicts that followed the food crisis decades; several, including Mukherjee, had served in the governments those verdicts made and unmade. The programme reads, on close inspection, less like a manifesto than like a memorandum from experience.

Two features of the text distinguish translation from rhetoric, and both are easy to miss in summary accounts. The first is its legalism. The employment commitment was not a scheme announcement but a promise of an Act, a statute that would create rights, duties, timelines and penalties, and the information and social security commitments were drafted in the same register. Legalism is expensive: it surrenders the discretion that governments prize, and a coalition surrenders it only when the commitment must be made credible to partners and constituencies who do not trust the government’s goodwill. The Left’s negotiators demanded statute because their outside position gave them no ministers to police a scheme; the rural constituencies’ organisations, several of which were already moving toward the advisory machinery described in the next section, demanded statute because they had watched schemes evaporate at the district office. The second feature is the text’s sectoral address. For a document governing a whole economy, the programme is strikingly uninterested in the sectors that had dominated the outgoing government’s publicity: there is no celebration of the technology enclave, no promise to the metropolitan consumer, no equivalent of India Shining’s mood music. Its centre of gravity is the rural household, its named anxieties are water, work, credit, food and school, and its fiscal promises run toward the districts that had moved in April and May. Readers who set the programme beside the first employment bill for comparison, a comparison made most legibly in a PowerPoint viewer that keeps the two texts’ clause structure intact, can watch the translation happening at the level of the sentence: the programme’s one hundred days become the bill’s household entitlement, its asset-creating works become the bill’s schedule of permissible works, its minimum wage becomes the bill’s wage clause. Interpretation had become drafting, and drafting is where interpretations are audited.

The National Advisory Council: Civil Society Inside the Policy Gate

A programme is a promise; the question of June 2004 was which machinery would convert the promise into bills a cabinet could carry. The answer, constituted by government order in June 2004 with Sonia Gandhi as chairperson, was the National Advisory Council, a body without constitutional or statutory basis whose membership joined retired officials, academics and civil society campaigners to advise the prime minister on the programme’s implementation. The council’s first-term records, preserved in its published communications to the government between 2004 and 2009, are the under-cited primary source this article was commissioned to use, and they repay the use, because they document a translation mechanism rather than a debating society. Within weeks of its constitution the council had taken up the two bills that headed the programme’s social agenda. Aruna Roy, the former Indian Administrative Service officer who had built the Mazdoor Kisan Shakti Sangathan in Rajasthan and led the campaign that won Rajasthan’s own information law in 2000, and Jean Drèze, the development economist who had worked on employment guarantee design since the Maharashtra precedent of the 1970s, carried drafts into the council’s early meetings. The employment guarantee draft that emerged from the council’s meeting of 19 August 2004 became the working basis of the bill the government introduced in Parliament that December.

Which two draft laws reached the Union Cabinet from the National Advisory Council in 2004?

The employment guarantee draft and the information law draft reached the Union Cabinet from the National Advisory Council in 2004. The council’s meeting of 19 August 2004 settled the employment draft’s architecture, and its members carried the information draft, built on the lapsed 2002 statute, through the same autumn’s deliberations.

The council’s significance for the sectoral argument is compositional. Its civil society members were not ornamentation drawn from the capital’s seminar circuit; they were the organisational residue of the very constituencies the verdict had moved. Roy’s organisation worked among agricultural labourers and marginal cultivators in Rajsamand district. The Right to Food Campaign, whose steering group overlapped with the council’s membership and whose state chapters had documented starvation deaths in the drought years, connected the council to the food security constituency of Rajasthan, Chhattisgarh and Orissa. M.S. Swaminathan, the agricultural scientist who chaired the National Commission on Farmers constituted in February 2004 and whose commission’s interim reporting through 2004 and 2005 documented the agrarian crisis in official prose, sat in the same advisory ecology. When the council drafted, the sector was in the room in a way it had never been inside a Congress Working Committee or a finance ministry budget meeting, and the design features that distinguish the resulting statutes, the job card, the muster roll open to inspection, the social audit conducted in the village assembly, bear the fingerprints of people who had spent a decade demanding exactly those instruments at exactly those sites. The conventional history of the period, organised around the personalities of Gandhi and Singh, treats the council as an expression of Gandhi’s power. The council’s own paper trail supports a different description: it was the institutional form in which the verdict’s constituencies were given drafting rights over the answer to their own demand.

The council also attracted the period’s sharpest institutional criticism, and the criticism belongs in the record at its strongest. Devesh Kapur and Pratap Bhanu Mehta, in their work on Indian public institutions collected in Public Institutions in India, published by Oxford University Press in 2005, and in their later editing of Rethinking Public Institutions in India in 2017, pressed the accountability problem that unelected and advisory bodies pose for a parliamentary system: authority exercised without electoral answerability must justify itself by the quality of its outputs, and even good outputs leave the democratic chain of sanction weakened at the link where the advice is given. Applied to the council, the argument runs that a body chaired by the ruling alliance’s leader, staffed partly by campaigners with programmes of their own, and reporting to a prime minister who did not chair it, blurred the location of decision in a way the Constitution’s cabinet scheme does not contemplate. Defenders of the council, including Roy in her published essays of the period, answered that the council advised and the cabinet decided, that every draft it produced faced Parliament unamended by any council privilege, and that the alternative location for the same drafting, inside ministries staffed by generalist officials, is no more electorally answerable and considerably less informed by the constituencies affected. The dispute is recorded here without a verdict of this article’s own beyond the evidentiary one: whatever the council’s constitutional merits, its paper trail is the best surviving record of the translation this article describes, and a history of the period that ignores the trail because the institution was irregular will misdescribe how the period’s central statutes acquired their shape.

The council’s first term, running to 2009, extended the same method across the programme’s remaining clauses: the forest rights legislation of 2006, the unorganised workers’ social security legislation of 2008, the food security drafting that occupied its later years and produced the council’s recommendations of 2010 and 2011 under its second term. Each exercise repeated the founding pattern, a constituency organisation’s draft, a council negotiation with ministry officials, a cabinet bill that diluted or disciplined the draft, and a public argument between the drafters and the government they advised. The pattern’s persistence across five years is itself evidence about 2004. Governments do not maintain an institution that embarrasses them, as the council repeatedly embarrassed the finance and rural development ministries, unless the institution performs a function the government cannot perform for itself. In the UPA’s case the function was legitimation by consultation: the coalition’s programme claimed to speak for constituencies the Congress organisation barely reached, and the council was the device by which those constituencies could be shown, in documentary form, to have been heard. The sectoral verdict had created a representational deficit at the heart of the government that won it. The council was the deficit’s institutional acknowledgment.

MGNREGA: The Employment Guarantee as Sectoral Translation

The National Rural Employment Guarantee Act, enacted in September 2005 and renamed for Mahatma Gandhi in October 2009, is the statute in which the sectoral reading becomes testable at the level of the clause. Its passage had not been smooth. The bill introduced in the Lok Sabha on 21 December 2004 by Raghuvansh Prasad Singh, the Minister of Rural Development, narrowed the council’s draft in respects that Roy and Drèze attacked in public within days: coverage was phased rather than universal at commencement, the guarantee’s enforcement machinery was thinner, and the wage and works provisions carried qualifications the drafters had resisted. The argument that followed through the first half of 2005, conducted in the press, in the council and in the parliamentary standing committee under the chairmanship of Kalyan Singh, is among the best documented drafting contests in Indian legislative history, and its outcome was a statute stronger in its entitlement core than the bill that had entered the House. Parliament passed the measure in August 2005; the President gave assent on 5 September 2005; and the Act commenced on 2 February 2006 across two hundred districts, launched at Bandlapalli village in Anantapur district, Andhra Pradesh, a choice of venue that closed the argument’s geographic circle in public: the guarantee began in the drought district whose grievance had opened the story.

Which minister piloted the employment guarantee bill in the Lok Sabha in 2004 and 2005?

Raghuvansh Prasad Singh, the Minister of Rural Development in the Manmohan Singh government, piloted the employment guarantee bill in the Lok Sabha. He introduced the measure on 21 December 2004, defended the government’s text against the advisory council’s wider draft through the standing committee stage, and carried its passage in August 2005.

The statute’s design answers the sectoral diagnosis point by point, and the correspondence is too systematic to be coincidence. The diagnosis held that the distressed constituency was the rural household dependent on seasonal wage work and rain-fed cultivation, exposed to a lean season in which no employer and no crop provided income. The Act therefore guarantees one hundred days of wage employment in a financial year to every rural household whose adult members volunteer for unskilled manual work, a household entitlement rather than an individual ration, pitched at the unit that actually pools rural income. The diagnosis held that the distress was worst where water had failed; the Act’s schedule of permissible works privileges water conservation, drought proofing, irrigation and land development, so that the wage bill doubles as an investment in the asset whose absence caused the distress. The diagnosis held that women agricultural labourers carried the crisis with the least bargaining power; the Act mandates equal wages for men and women and provides that at least one third of beneficiaries shall be women, and the programme’s recorded participation of women rose well beyond that floor in the southern states within its first years. The diagnosis held that the old works programmes had leaked through falsified muster rolls administered in secret; the Act builds its enforcement from the job card held by the household rather than the office, the dated application that triggers a legal duty to provide work within fifteen days or pay an unemployment allowance, the sixty to forty ceiling on the material share of works costs that protects the wage component, and the social audit conducted in the gram sabha with records open to any resident. Each feature had been demanded, in those terms, by the campaigns of the preceding decade, in Rajasthan’s public hearings of the 1990s and in the Right to Food Campaign’s submissions of 2001 to 2004. The statute is those demands wearing legal dress.

The renaming of October 2009, which attached Gandhi’s name to a statute the campaigns had built, and the programme’s expansion to all rural districts by April 2008, complete the first term’s arc, and the assessments that followed supply the adjudication this article owes its readers on the translation’s quality. The Comptroller and Auditor General’s performance audit of 2013 documented wage delays, incomplete works and administrative thinness across states; the Mihir Shah committee’s revision of the operational guidelines in 2012 had already conceded the design’s dependence on state administrative capacity that the Act could command but not create. Devesh Kapur’s work on state capability, and Pratap Bhanu Mehta’s essays on the gap between India’s legislative ambition and its administrative plumbing, collected across the 2010s, locate the statute’s uneven record in that dependence rather than in its entitlement structure. None of this retrospective evidence alters the historical claim, which is about what the verdict demanded and what the government of 2004 to 2005 understood itself to be answering. A translation can be faithful and still be executed unevenly by the machinery available. The faithfulness is established by the clause-level correspondence between the sectoral diagnosis and the statutory text. The unevenness belongs to a different argument, about the Indian state’s capacity, which the verdict of 2004 did not settle and no single statute could.

Information, Health and Forest Rights: The Wider Rights Architecture

The employment guarantee did not travel alone, and its companions in the legislative sequence of 2005 and 2006 show the sectoral translation widening from a wage instrument into an architecture. The Right to Information Act, passed by Parliament in May 2005, assented to on 15 June 2005 and brought into force on 12 October 2005, replaced the weak Freedom of Information Act of 2002, which the Vajpayee government had enacted and never notified. The new statute gave any citizen the right to demand records from public authorities against a thirty-day reply deadline, created independent information commissions at the Centre and in the states, and placed the burden of refusal on the authority withholding the record. Its genealogy is the genealogy of the sectoral verdict in miniature. The demand for information had been forged in Rajasthan in the 1990s, where the Mazdoor Kisan Shakti Sangathan’s public hearings in Devdungri and across Rajsamand district exposed wage theft on drought relief works by the simple method of reading official muster rolls aloud to the labourers named in them; the state law of 2000 followed, and Aruna Roy’s presence in the National Advisory Council in 2004 carried the demand from the hearing ground to the drafting table. The connection between the information statute and the employment statute is functional rather than rhetorical: a guarantee administered through panchayat records is enforceable only if the records are open, and the drafters of 2004 and 2005 treated the two bills as one system in two statutes, the entitlement and the instrument for policing it.

The National Rural Health Mission, launched by Manmohan Singh on 12 April 2005, carried the translation into the sector where rural India’s deficit was measured in lives rather than wages. The mission’s design targeted the eighteen states with the weakest health indicators, most of them rural and northern, and built its delivery around a new village-level worker, the Accredited Social Health Activist, a woman resident of the village she served, trained and incentivised to bring pregnant women to institutional deliveries, children to immunisation and patients to the public system. Its companion scheme, Janani Suraksha Yojana, paid cash assistance for institutional childbirth, and the mission’s flexible financing to district and facility committees broke with the rigid line-item budgeting that had starved primary health centres of untied funds. The mission belongs in this article because its constituency is the employment guarantee’s constituency in another register: the rural household whose poverty was compounded by a health system that charged it, in fees, travel and lost wages, for every illness. The Common Minimum Programme’s health spending commitment supplied the mission’s warrant, and the mission’s existence within the government’s first year, designed and launched before the employment bill had completed its passage, shows the translation proceeding on several fronts at once rather than as a single flagship gesture.

The third companion statute, the Scheduled Tribes and Other Traditional Forest Dwellers (Recognition of Forest Rights) Act, passed in December 2006 and assented to on 29 December 2006, extended the architecture to the constituency whose dispossession was the oldest in the republic’s ledger: the forest-dwelling communities whose customary rights the colonial forest settlements had extinguished and whose post-independence treatment the state’s own committees had repeatedly indicted. The Act recognised individual rights to forest land under cultivation and community rights over forest produce and governance, vesting the initial determination in the gram sabha. Its passage required the government to overcome resistance inside its own forest bureaucracy and a vigorous campaign by conservation organisations that feared, and said in print through 2005 and 2006, that recognition would fragment the protected forest estate; the tribal organisations that had drafted toward the statute for a decade answered that the estate’s protection had been built on the criminalisation of its oldest inhabitants. The dispute is noted here because it marks the boundary of the council method: the forest rights bill moved through a Group of Ministers and a parliamentary joint committee rather than through the advisory council’s drafting alone, and its two-year passage, against the employment bill’s eight months, measures how far the sectoral mandate’s authority extended. Where the verdict’s core constituency and the programme’s text aligned, translation was swift. At the architecture’s edge, where organised interests contested the constituency’s claim, the same government moved at the older speed of Indian legislation.

The Long Translation: Food Security, and the Limits of the Reading

The food security legislation completed the programme’s arc on a timescale that tests the sectoral reading against delay. The Common Minimum Programme had promised a strengthened public distribution system and movement toward food security; the National Advisory Council’s drafting on a food bill ran through 2010 and 2011, producing recommendations for near-universal coverage that the government’s bill of December 2011 narrowed to a targeted entitlement; the National Food Security Ordinance followed on 3 July 2013, Parliament enacted the statute in September 2013, and presidential assent came on 10 September 2013, with the Act deemed in force from 5 July 2013. The enacted text entitled up to three quarters of the rural population and half of the urban population to five kilograms of foodgrains per person per month at prices fixed in the statute, made the eldest woman of the household its statutory head for ration purposes, and attached maternity entitlements and children’s meal guarantees through the anganwadi and school systems. The distance between the council’s universal draft and the enacted targeting, and the nine years between promise and statute, record the resistance the translation met from fiscal authorities and state governments once the first term’s momentum spent itself. The sectoral reading does not require the translation to have been swift or complete. It requires the agenda to have been set by the verdict’s content, and the food statute’s nine-year gestation inside the programme’s framework, against finance ministry opposition recorded at every stage, is stronger evidence of agenda-setting than a quick statute would have been: the commitment survived two elections, a change of council membership and a decade of fiscal argument because it had been written into the founding text of May 2004 as the verdict’s due.

The limits of the reading should be stated with the same plainness as the reading itself, and three limits bind it. The first is the urban evidence. Yogendra Yadav’s constituency analysis in The Elusive Mandate of 2004 recorded that the National Democratic Alliance’s seat losses were heaviest in urban and metropolitan constituencies, where its tally fell far more sharply than in rural seats, a finding the sectoral reading must absorb rather than deny. The absorption runs through the alliance mechanism: metropolitan seats in Hyderabad, Chennai, Delhi and Mumbai changed hands because state alliances re-sorted and state incumbencies collapsed, and the voters of India’s cities in 2004 were voting inside state arguments, on state records, in constituencies whose economies were bound to their agrarian hinterlands by migration, remittance and food prices. The urban seat fall is a fact about where the electoral system registered the movement, not a refutation of the movement’s sectoral content. The second limit is turnout. The National Election Study’s participation analysis, in the work of Suhas Palshikar and Sanjay Kumar on the 2004 study, records rural turnout falling relative to 1999 rather than surging; the sectoral verdict was delivered by vote switching and alliance pooling among those who voted, not by a countryside rising in exceptional numbers, and any romance of rural revolt is excluded by the study the reading itself relies on. The third limit is the Congress’s own weakness: a party that lost national vote share in victory, that stagnated in Uttar Pradesh and collapsed in Karnataka, was an improbable vessel for a national mandate, and the reading defended here is precisely that there was no national mandate for it to carry. There was a set of state verdicts with a common sectoral thread, a programme that priced the thread, and a legislative sequence that honoured the price. That is a smaller claim than the mythology of 2004 usually makes, and it is the claim the evidence supports.

Adjudication can therefore be rendered. The rural-backlash reading, in the form in which it entered the textbooks and the reference summaries, asserts a generic national rejection of the Bharatiya Janata Party driven by rural anger at India Shining. The Election Commission record contradicts its geography, since the party held its rural heartland states; the National Election Study contradicts its mechanism, since the movement ran through switching and pooling rather than turnout; the state record contradicts its agency, since the decisive destructions were of regional incumbents in Andhra Pradesh and Tamil Nadu and of a Congress incumbent in Karnataka; and the legislative record contradicts its content, since the government formed from the verdict legislated for a named sector with named instruments in a sequence traceable from the drafting table of May 2004 to the statute book of 2005 and 2006. The sectoral reading accounts for each of those records without remainder, including the two records, Karnataka and the urban seat fall, that a weaker thesis would have to hide. The 2004 verdict was a sectoral verdict: delivered by agrarian constituencies through state arenas, converted by alliance design, and translated by a coalition that wrote its interpretation into law while the interpretation could still be checked against the result. Its lesson for the reading of Indian elections is the lesson Yadav’s elusive mandate taught at the time and the statute book confirmed afterward. A verdict is not what the winning side says it was on counting night. A verdict is what the seat arithmetic, the constituency record and the first year’s legislation, read together, can be shown to have demanded.

The 2009 Check: A Second Verdict Read Against the First

A reading of an election earns confidence when it survives contact with the next one, and the general election of April and May 2009 supplies that contact. The Election Commission record for 2009 shows the Congress rising to 206 seats, its best tally since 1991, the Bharatiya Janata Party falling to 116, and the United Progressive Alliance returning to office with a larger core and a smaller dependence on outside backing than in 2004. The CSDS National Election Study of 2009, analysed by Yogendra Yadav and Suhas Palshikar in their post-poll writing of that year, recorded an incumbent advantage operating state by state rather than a national wave, the same structure the 2004 analysis had found beneath a change of government. For this article’s argument the pertinent evidence is geographic and programmatic. The Congress’s largest gains came in Andhra Pradesh, where it won thirty-three of forty-two seats under the same Y.S. Rajasekhara Reddy government that had made the free power and irrigation promises of 2004 and had spent five years delivering them, and in Rajasthan and Haryana, where state governments had aligned themselves with the employment guarantee and the farm loan waiver of 2008. The loan waiver itself, announced in the Union budget of 29 February 2008 by P. Chidambaram at a cost of roughly sixty thousand crore rupees to the exchequer, was the programme politics of 2004 extended by other means: a direct, named transfer to the cultivating sector, defended by its authors in the language of agrarian distress and attacked by its critics, among them the Comptroller and Auditor General in a later audit of its implementation, as fiscal populism. The attack and the defence agree on the fact that matters here, which is the sector addressed.

Karnataka in 2009 repeated its instructional role. The Bharatiya Janata Party, governing the state since May 2008 under B.S. Yeddyurappa, held nineteen of twenty-eight Lok Sabha seats, and the Congress held six, a distribution that once again tracked state incumbency and organisation rather than any national preference for or against the Delhi government. Bihar likewise re-sorted itself in 2009 around Nitish Kumar’s state government, whose Janata Dal (United) and Bharatiya Janata Party alliance won thirty-two of forty seats on a record of state administration, reversing the state’s 2004 alignment with the same indifference to Delhi’s composition that the 2004 reversal had shown. The pattern across the two elections is the sectoral reading’s signature. Indian voters in this period rewarded and punished state-level managements of the agrarian and welfare economy, the Union government’s survival depended on its programme being legible as an answer to that economy’s demands, and the party that had written the answer into statute in 2005 harvested the incumbency of 2009 in the states where its own governments had delivered it. A generic anti-Bharatiya Janata Party mood explains none of the 2009 distributions, since the mood would have had to reverse itself in Karnataka and Bihar while intensifying in Andhra Pradesh. The sectoral mechanism, mediated by incumbency and delivery, explains all three.

Frequently Asked Questions

Q: What was the 2004 UPA verdict?

The 2004 UPA verdict was the result of the general election counted on 13 May 2004, in which the Congress emerged as the largest party with 145 seats and the Bharatiya Janata Party fell to 138, ending Atal Bihari Vajpayee’s government. The Congress then assembled the United Progressive Alliance from its pre-poll allies and post-poll partners, reaching roughly 217 to 222 seats depending on how alliance membership is classified, and took office on 22 May 2004 with the Left parties, holding about sixty seats, sustaining the government from outside the cabinet. The verdict’s distinctive feature was its narrowness at the level of votes: the Congress won its seats on 26.53 percent of the national vote, a smaller share than it had taken in defeat in 1999, so the result is best understood as an assembly of state verdicts rather than a national wave.

Q: What is the India Shining campaign?

The India Shining campaign was the publicity exercise mounted by the National Democratic Alliance government from late 2003 into the election of 2004, presenting India as an economy in confident ascent, with high growth, expanding mobile telephony, new highways and rising foreign investment as its exhibits. The campaign began as government advertising and became the ruling alliance’s election atmosphere, reinforced by the slogan of a feel-good factor after the Bharatiya Janata Party’s state election victories of December 2003. Its political defect, identified by the Congress campaign under Sonia Gandhi and by commentators through 2004, was the distance between the advertised mood and the recorded condition of the agrarian economy, where drought, indebtedness and falling farm incomes had marked the years from 2000 onward. The slogan’s failure became the conventional shorthand for the verdict, although the seat evidence shows the defeat was assembled state by state.

Q: Why did the BJP lose in 2004?

The Bharatiya Janata Party lost in 2004 through a compound of alliance losses, state incumbencies and sectoral grievances rather than through a collapse of its own vote, which fell from 23.75 percent in 1999 to 22.16 percent in 2004 on the Election Commission count. Its 1999 partners deserted or collapsed in the decisive states: the Telugu Desam Party fell from twenty-nine seats to five in Andhra Pradesh, and the Dravida Munnetra Kazhagam, Pattali Makkal Katchi and Marumalarchi Dravida Munnetra Kazhagam crossed to the Congress-led alliance in Tamil Nadu, where the Bharatiya Janata Party’s new partner, the All India Anna Dravida Munnetra Kazhagam, won nothing. The early election gamble of February 2004, taken after favourable state results, exposed the alliance to agrarian distress in the Deccan and the south before any national campaign could price it. The party held its heartland states, which is why analysts describe the defeat as an alliance and geography failure.

Q: Who became Prime Minister after 2004?

Manmohan Singh became Prime Minister after the 2004 election, sworn in by President A.P.J. Abdul Kalam on 22 May 2004. Singh, an economist who had served as finance minister from 1991 to 1996 under P.V. Narasimha Rao and was a Rajya Sabha member from Assam, was named after Sonia Gandhi, elected leader of the Congress Parliamentary Party on 15 May 2004, declined the office on 18 May 2004 despite holding the letters of backing needed to claim it. Kalam appointed Singh on 19 May 2004 after Gandhi presented the alliance’s backing and named him as its head. Singh led the United Progressive Alliance government for two full terms, from 2004 to 2014, the longest tenure of any Indian prime minister who had not led his party into the election that installed him.

Q: What is MGNREGA?

MGNREGA is the Mahatma Gandhi National Rural Employment Guarantee Act, enacted in September 2005 as the National Rural Employment Guarantee Act and renamed in October 2009. The statute guarantees one hundred days of wage employment in each financial year to every rural household whose adult members volunteer for unskilled manual work, on public works that privilege water conservation, drought proofing and rural assets. Its enforcement design includes a job card issued through the gram panchayat, a legal duty to provide work within fifteen days of application or pay an unemployment allowance, equal wages for men and women, a requirement that at least one third of beneficiaries be women, a ceiling on the material share of works costs, and social audits conducted in the village assembly. The Act commenced on 2 February 2006 in two hundred districts and extended to all rural districts by April 2008.

Q: What is the Common Minimum Programme of UPA?

The Common Minimum Programme of the United Progressive Alliance, adopted as the National Common Minimum Programme and released on 27 May 2004, was the negotiated governing text of the coalition that took office after the 2004 election, drafted by the Congress and its allies with the Left parties whose outside backing sustained the government. Its commitments included a legal guarantee of at least one hundred days of employment on asset-creating public works, enhanced public investment and credit for agriculture, education spending rising toward six percent of gross domestic product, health spending rising toward two to three percent, universalisation of the Integrated Child Development Services, a strengthened public distribution system and social security for unorganised workers. Finance Minister P. Chidambaram organised the Union budget of 8 July 2004 around the programme, calling it the government’s guiding light, and its clauses supplied the drafting agenda for the employment, information, health and food statutes of the following years.

Q: Did the Congress win the 2004 election?

The Congress won the 2004 election in the sense that it emerged as the largest party, with 145 seats against the Bharatiya Janata Party’s 138, and led the government that took office on 22 May 2004. It did not win a majority, falling 127 seats short of the 272 mark, and it did not win a larger share of the vote than in its 1999 defeat: its 26.53 percent in 2004 was below the 28.30 percent it had polled five years earlier. The victory was therefore coalitional and geographic. Pre-poll allies such as the Rashtriya Janata Dal, the Dravida Munnetra Kazhagam and the Nationalist Congress Party, post-poll partners, and the Left parties’ outside backing converted the Congress plurality into a governing majority. Describing 2004 as a Congress win is accurate about office and misleading about mandate, which is why the result is usually analysed state by state.

Q: What was the National Advisory Council, and who chaired it in 2004?

The National Advisory Council was an advisory body constituted by government order in June 2004 to advise the prime minister on implementing the National Common Minimum Programme, and it was chaired in 2004 by Sonia Gandhi, president of the Congress and chairperson of the United Progressive Alliance. Its membership joined retired officials and academics to civil society campaigners, among them Aruna Roy of the Mazdoor Kisan Shakti Sangathan and the economist Jean Drèze, and its early meetings produced the drafts behind the employment guarantee and information statutes of 2005. The council’s published communications to the government between 2004 and 2009 record its work on forest rights, unorganised workers’ social security and food security drafting. Critics questioned the accountability of an unelected body with drafting influence; its defenders answered that it advised in the open and that cabinet and Parliament decided.

Q: What did the Right to Information Act of 2005 change about access to government records?

The Right to Information Act of 2005 replaced the Freedom of Information Act of 2002, which had been enacted under the Vajpayee government and never brought into force, and it changed access to government records from an administrative concession into a citizen’s enforceable right. Passed by Parliament in May 2005, assented to on 15 June 2005 and in force from 12 October 2005, the Act entitled any citizen to demand information from public authorities against a reply deadline of thirty days, created independent information commissions for the Union and the states with power to order disclosure and impose penalties, required proactive publication of official information, and placed the burden of justifying a refusal on the authority withholding the record. Its earliest heavy use came from the constituencies that had campaigned for it, including applicants policing wage and ration records under the rural programmes of the period.

Q: What was the National Rural Health Mission launched in April 2005?

The National Rural Health Mission was the Manmohan Singh government’s flagship rural health programme, launched on 12 April 2005 to rebuild public health delivery in the countryside, with concentrated attention on eighteen states whose health indicators were weakest. Its best known instrument was the Accredited Social Health Activist, a woman resident of the village she served, selected from that village and trained to mobilise households toward institutional childbirth, immunisation and public treatment. The mission’s companion scheme, Janani Suraksha Yojana, paid cash assistance to mothers for institutional deliveries, and its financing gave district societies and facility committees untied funds that the older line-item system had denied them. The mission answered the health clause of the Common Minimum Programme of May 2004, which had committed the government to raise public health spending, and it addressed the same rural households the employment guarantee was designed to reach.

Q: How did Andhra Pradesh vote in the 2004 Lok Sabha election?

Andhra Pradesh voted in 2004 to dismantle the strongest state machine in the National Democratic Alliance. The Election Commission record shows the Congress winning twenty-nine of the state’s forty-two seats, the Telangana Rashtra Samithi five, the All India Majlis-e-Ittehadul Muslimeen one and the two Communist parties one each, against five seats for the Telugu Desam Party and none for the Bharatiya Janata Party, which had held seven in 1999 alongside the Telugu Desam Party’s twenty-nine. The simultaneous assembly election gave the Congress 185 of 294 seats, and N. Chandrababu Naidu resigned as chief minister on 11 May 2004. The campaign had been organised by Y.S. Rajasekhara Reddy around farm power, irrigation and loan relief after his 2003 padayatra through the drought districts, and the Congress-led alliance ran strongest, on the CSDS state sample, among agricultural labourers and the poorest respondents.

Q: How did Tamil Nadu vote in the 2004 Lok Sabha election?

Tamil Nadu voted in 2004 for a complete transfer of its delegation. The Democratic Progressive Alliance led by M. Karunanidhi’s Dravida Munnetra Kazhagam won all thirty-nine seats, divided among the Dravida Munnetra Kazhagam with sixteen, the Congress with ten, the Pattali Makkal Katchi with five, the Marumalarchi Dravida Munnetra Kazhagam with four and the Communist Party of India and the Communist Party of India (Marxist) with two each. The rival combination of the All India Anna Dravida Munnetra Kazhagam of J. Jayalalithaa, the state’s incumbent chief minister, and the Bharatiya Janata Party won no seat, reversing the 1999 pattern in which a Dravida Munnetra Kazhagam-led alliance inside the National Democratic Alliance had held twenty-six seats. The verdict punished the state government’s record on drought relief, farm power and welfare entitlements, and it delivered the largest single state bloc inside the incoming coalition’s majority.

Q: Why does Karnataka complicate the sectoral reading of the 2004 verdict?

Karnataka complicates the sectoral reading because the state’s agrarian distress, as severe as any in the south after four years of drought, punished the Congress rather than the Bharatiya Janata Party. The Congress, which governed the state under S.M. Krishna and held eighteen of twenty-eight Lok Sabha seats in 1999, fell to eight in 2004, and the Bharatiya Janata Party rose from seven seats to eighteen, with the Janata Dal (Secular) holding two. The complication disciplines the thesis rather than defeating it. Distress in 2004 moved against whichever party held the state government answerable for water, power and prices, and in Karnataka that party was the Congress, whose celebrated urban administration in Bengaluru stood accused by its opponents of neglecting the drying northern districts. A reading based on incumbency and sector predicts the Karnataka result; a reading based on a national anti-Bharatiya Janata Party mood cannot.

Q: How many seats did the Left parties win in 2004, and why did their outside support matter?

The Left parties won fifty-nine seats in 2004 on the Election Commission count, built from forty-three for the Communist Party of India (Marxist), ten for the Communist Party of India, three for the Revolutionary Socialist Party and three for the All India Forward Bloc, their strongest combined performance in the republic’s history. Their outside backing mattered because it converted the Congress-led alliance’s plurality into a working majority without bringing the Left into the cabinet: the United Progressive Alliance core of roughly 217 to 222 seats, added to the Left’s fifty-nine, placed the government beyond the reach of a confidence challenge. The price of the arrangement was written into the National Common Minimum Programme of 27 May 2004, whose employment, agriculture and social spending clauses carried the Left’s conditions, and the arrangement replicated, at national scale, the outside backing model the Congress itself had extended to the United Front governments of 1996 to 1998.

Q: What role did Aruna Roy play in the employment guarantee campaign?

Aruna Roy, a former Indian Administrative Service officer who resigned in 1975 and built the Mazdoor Kisan Shakti Sangathan in Rajasthan from 1990, carried the employment guarantee from the hearing ground into the statute book. Her organisation’s public hearings of the 1990s, at which official muster rolls for drought relief works were read aloud to the labourers named in them, forged both the information campaign that won Rajasthan’s state law of 2000 and the design demands, job cards, open records, social audits, that later structured the national guarantee. Appointed to the National Advisory Council in 2004, she and Jean Drèze pressed the council’s wider employment draft upon the government, and when the bill introduced on 21 December 2004 narrowed that draft, Roy criticised the dilution in public and campaigned with the Right to Food network for the stronger text Parliament enacted in August 2005.

Q: What did the Forest Rights Act of 2006 recognise?

The Forest Rights Act of 2006, formally the Scheduled Tribes and Other Traditional Forest Dwellers (Recognition of Forest Rights) Act, passed in December 2006 and assented to on 29 December 2006, recognised rights over forest land and produce that colonial forest settlements and post-independence administration had treated as encroachments. The Act recognised individual rights to forest land held for habitation and cultivation, community rights to minor forest produce, grazing, water bodies and traditional seasonal access, and the right of communities to protect and manage their customary forest. It vested the initiation of claims in the gram sabha, the village assembly, with appeals through subdivisional and district committees, and it set eligibility conditions that tribal claimants and other traditional forest dwellers had to meet through evidence of residence and dependence. Implementation rules followed in January 2008, and the statute’s recognition agenda formed part of the rights architecture the 2004 programme had promised.

Q: How did the National Food Security Act extend the UPA policy programme in 2013?

The National Food Security Act extended the United Progressive Alliance’s policy programme by converting the public distribution commitments of the Common Minimum Programme of May 2004 into a statutory food entitlement. Enacted by Parliament in September 2013 after an ordinance of 3 July 2013, assented to on 10 September 2013 and deemed in force from 5 July 2013, the Act entitled up to seventy-five percent of the rural population and fifty percent of the urban population to five kilograms of foodgrains per person per month at prices fixed in the statute, three rupees a kilogram for rice, two for wheat and one for coarse grains. It named the eldest woman of each eligible household as its statutory head for ration purposes, attached maternity entitlements, and guaranteed children’s meals through the anganwadi and school meal systems. The Act’s drafting had run through the National Advisory Council from 2010 onward.

Q: Why did Sonia Gandhi decline the prime ministership in May 2004?

Sonia Gandhi declined the prime ministership in May 2004 in the explanation she gave the Congress Parliamentary Party on 18 May 2004: she said she had listened to her inner voice, that power had never been her aim, and that she would not take an office she had not sought for herself. The context her contemporaries recorded included a Bharatiya Janata Party campaign against her Italian birth, in which Sushma Swaraj had threatened in April 2004 to shave her head if Gandhi took the office, and litigation threats against her eligibility that constitutional lawyers of the period dismissed. A later account by Natwar Singh, published in 2014, claimed her son Rahul Gandhi’s fear for her safety decided the refusal; Gandhi never adopted that account, and the motive remains disputed between the public statement of 2004 and the retrospective claims. The institutional result is undisputed: Manmohan Singh became prime minister and Gandhi led the alliance.

Q: What did Yogendra Yadav mean by calling the 2004 mandate elusive?

Yogendra Yadav, the Centre for the Study of Developing Societies analyst who directed the reading of the National Election Study of 2004, called the mandate elusive in his Economic and Political Weekly essay of December 2004, The Elusive Mandate of 2004, to warn against treating the seat result as a single national instruction. His argument rested on the electoral system’s amplification: modest and contradictory vote movements, in a federation of state party systems, had produced a dramatic seat reversal, with both national parties losing vote share even as the Congress gained thirty-one seats and the Bharatiya Janata Party lost forty-four. A mandate assembled from state verdicts delivered for different reasons could not honestly be summarised as one message, whether the summary claimed a rejection of economic reform, a rural revolt or a secular restoration. The elusiveness was a property of the evidence, and Yadav’s remedy was disaggregation to the state and the constituency.

Q: How did the 2002 Gujarat violence shape voting and alliance choices in 2004?

The Gujarat violence of February and March 2002, in which more than a thousand people, most of them Muslims, were killed after the Godhra train burning of 27 February 2002, shaped the 2004 election chiefly by setting the terms on which alliances could be built and votes coordinated. The Communist Party of India (Marxist), in its Central Committee review adopted between 30 July and 1 August 2004, listed the defeat of the Bharatiya Janata Party and the installation of a secular government as its first objectives, and the Left’s sixty seats were committed to a Congress-led government on that basis. Parties such as the Dravida Munnetra Kazhagam and the Lok Janshakti Party crossed away from the National Democratic Alliance without paying a penalty among their own electorates, and Muslim voters in mixed constituencies coordinated behind the strongest local alternative. The effect was bounded by state arenas: the Bharatiya Janata Party retained fourteen of Gujarat’s twenty-six seats.