The 1996 Vajpayee thirteen days are the most instructive failure in the history of Indian coalition politics. A government that passed no law, presented no budget and never faced the vote that would have destroyed it nonetheless redrew the strategic map of the party system, because the manner of its fall announced the method of its party’s return. Atal Bihari Vajpayee took oath as Prime Minister on 16 May 1996 at the head of the single largest party in a hung Lok Sabha, resigned thirteen days later on 28 May 1996 after a confidence debate he chose to close with a doctrine instead of a bargain, and watched H.D. Deve Gowda take oath in his place on 1 June 1996. The standard reference treatment files the episode as a footnote, a ministry too brief to matter. The argument developed across this article is the opposite one: the thirteen days mattered for what they signalled rather than for what they accomplished, and the resignation speech that ended them was the foundation document of the coalition that took office in March 1998.

The 1996 Vajpayee Thirteen Days in the Standard Telling

The standard telling is accurate in every particular and wrong in its conclusion. Elections to the Eleventh Lok Sabha were held on 27 April, 2 May and 7 May 1996. The Election Commission of India’s statistical report on the 1996 general election records the outcome that every later event obeyed: the Bharatiya Janata Party won 161 seats on 20.29 percent of the vote, the Indian National Congress won 140 seats on 28.80 percent, the Janata Dal won 46, the Communist Party of India (Marxist) won 32, and the remainder scattered across regional parties whose individual tallies were small and whose collective weight was decisive. No party stood within reach of a majority. President Shankar Dayal Sharma invited the leader of the largest party to form a government, Vajpayee took oath on 16 May 1996, the government failed to gather the numbers, Vajpayee resigned on 28 May 1996 rather than lose a confidence vote, and the United Front government of H.D. Deve Gowda followed on 1 June 1996. Textbook chapters give the ministry a paragraph. The standard biographical treatments give it colour, the drama of a poet statesman accepting a doomed commission with grace, and move on to Pokhran and the fuller tenures.

What the footnote treatment cannot explain is the sequel. Within twenty-two months of the resignation, on 19 March 1998, Vajpayee took oath again, at the head of a coalition of more than twenty parties organised under a negotiated common programme, with regional parties that had refused him in 1996 queued inside or behind his government. A footnote does not produce a sequel of that kind. Something happened inside the thirteen days, and inside the twenty-two months that followed them, that converted a party treated as untouchable into the natural centre of a governing alliance. That conversion, its mechanism and its documentary record, is the subject of the sections that follow.

Which party won the largest vote share in the 1996 election?

The Congress won the largest vote share in 1996 and finished second in seats. The Election Commission’s statistical report credits the Congress with 28.80 percent of the valid vote and 140 seats, against the BJP’s 20.29 percent and 161 seats, an early and stark demonstration that the post-Mandal party system had begun to reward geographical concentration over national spread.

Thirteen days also requires a calibration the footnote treatment rarely supplies, because brevity in Indian prime ministerial history is not itself a distinction. Charan Singh held the office from 28 July to 20 August 1979, resigned when the Congress withdrew its backing, and never faced the Lok Sabha at all, governing on in a caretaker capacity for months while the country moved to the election of January 1980. Chandra Shekhar governed from November 1990 to March 1991 at the pleasure of the Congress’s outside backing and resigned when that backing turned conditional. Gulzarilal Nanda’s two caretaker interludes of 1964 and 1966 were shorter than Vajpayee’s tenure and are remembered, correctly, as pure procedure. What distinguishes May 1996 from that company is not the brevity but the content poured into it: a contested confidence debate fought to its final hour, a resignation announced on the floor rather than conveyed by letter, and a doctrine of future conduct stated on the record by the resigning prime minister. Charan Singh’s fall produced an election and a restoration. Vajpayee’s produced a strategy. The calibration matters because the structural reading does not ask the reader to marvel at a short tenure; it asks the reader to notice that a short tenure was used, deliberately and in public, as an instrument, which is a different and rarer thing.

The disjunction between votes and seats in that report deserves a pause, because it frames everything else. The BJP’s support was concentrated in the northern and western states, where first-past-the-post converted pluralities into seats with brutal efficiency. The Congress vote was spread across nearly every state, deep enough to finish a respectable second almost everywhere and shallow enough to win almost nowhere outside its surviving strongholds. The Janata Dal, the Samajwadi Party, the Bahujan Samaj Party and the Dravida parties drew intense support from defined social constituencies within defined territories. The Eleventh Lok Sabha was therefore not a parliament of three national parties but a parliament of one declining national party, one rising regionalised party with national ambitions, and a tier of state parties whose seats no government could assemble a majority without. The thirteen days were the first time that structure was put to a formal test, and the structure won.

The Verdict Assembled: How the Election of 1996 Was Fought

The verdict that produced the thirteen days was assembled across a campaign in which every major party carried a disabling weakness into the contest. The Congress fought the election as the incumbent of five bruising years: the Rao government had survived its term, but the final year had been consumed by scandal, capped by the Central Bureau of Investigation’s chargesheets in the Jain hawala case in early 1996, which named politicians across the spectrum in a diary of alleged payoffs and forced resignations and disqualifications from public life across the party divide. The BJP fought under a shadow of the same case. L.K. Advani, the party’s president and its most formidable campaigner, was among those named, and Advani announced that he would not enter the Lok Sabha until the charge was cleared, a stand that removed the party’s second figure from the electoral field and concentrated the entire campaign on Vajpayee, who led it as the declared prime ministerial face. The Janata Dal fought as a federation of state satraps already drifting apart, its national leadership unable to impose a common campaign on Lalu Prasad Yadav in Bihar, Biju Patnaik in Orissa or the party’s Karnataka unit under Deve Gowda. The Left fought to defend its two citadels of West Bengal and Kerala and to hold its national bloc intact.

The results converted those weaknesses into a precise distribution of strength. The BJP’s 161 seats were won overwhelmingly in the northern and western states, with the party entrenched in Uttar Pradesh, Madhya Pradesh, Rajasthan, Gujarat, Maharashtra in harness with the Shiv Sena, and Delhi. The Congress’s 140 were the residue of a national presence: the party finished first or second in more states than any rival and converted that presence into seats at a rate that punished breadth. The regional parties converted depth: the Samajwadi Party’s 17 seats in Uttar Pradesh, the Bahujan Samaj Party’s 11, the Dravida Munnetra Kazhagam’s 17 in Tamil Nadu, the Tamil Maanila Congress’s 20 in the same state, the Telugu Desam’s 16 in Andhra Pradesh. Two features of the map governed everything that followed. The BJP had no presence worth the name in the south outside Karnataka’s handful or in the east outside its Orissa and Bihar pockets, so the party’s path to 272 ran through parties rooted in regions where it could not campaign. And no two of the regional parties answered to a common command, so the path had to be walked party by party, at prices quoted party by party. The thirteen days were the BJP’s first attempt to walk it, undertaken in public, against a deadline, with the whole country watching the meter run.

The Post-Mandal Arithmetic Behind the Claim

Vajpayee’s invitation to form a government in May 1996 was a product of arithmetic that had been assembling since 1989. The implementation of the Mandal Commission’s recommendations by the V.P. Singh government in August 1990, reserving 27 percent of central government posts for the Other Backward Classes, reorganised electoral competition in the Hindi belt around caste constituencies that the Congress had once held inside a single catch-all coalition. The post-1989 verdict’s structural rupture is the foundation of the story: the 1989 election broke the Congress system, ended four decades of single-party dominance and installed a tripolar contest in the northern states among the Congress, the BJP and the parties of the Janata parivar. Through the early 1990s that tripolar contest hardened. The Samajwadi Party was founded in 1992 on a Yadav and Muslim base in Uttar Pradesh. The Bahujan Samaj Party converted Kanshi Ram’s organisation into eleven Lok Sabha seats in 1996. The Janata Dal fragmented state by state into successors that guarded their territories jealously.

The BJP’s rise ran on a parallel track. From 85 seats in 1989 and 120 in 1991, powered by the Ram Janmabhoomi mobilisation and by the Congress’s organisational decay, the party reached 161 in 1996 and the position of single largest party for the first time. The critical point about the 161 is what kind of majority-building material it was. The BJP’s seats sat atop a social coalition that stopped at the boundaries of its own mobilisation: upper castes, trading communities and the urban middle classes of the north and west, with thin presence in the south and east. The parties holding the remaining seats represented constituencies, the backward-caste parties of the Gangetic plain, the Dravida parties of Tamil Nadu, the Telugu Desam of Andhra Pradesh, whose voters had been mobilised in substantial part against the politics the BJP was taken to represent. Post-Mandal arithmetic could produce a non-Congress, non-Janata governing bloc for the first time in 1996, because the BJP had become too large to exclude from the first claim on power. The same arithmetic guaranteed that the claim, once made, would have to be redeemed in a currency the BJP had not yet learned to mint: negotiated partnership with parties whose social bases were not its own. That is the renegotiation at the heart of the coalition-era framework, and May 1996 was its first formal examination.

The tripolar structure had a geography, and the geography explains why no national formula could dissolve it. In Uttar Pradesh, the contest ran among the BJP, Mulayam Singh Yadav’s Samajwadi Party and the Bahujan Samaj Party of Kanshi Ram and Mayawati, three organisations drawing on social blocs, upper castes, backward Yadavs and Muslims, and Dalits, whose mutual antagonisms were the state’s political grammar. In Bihar, the Janata Dal of Lalu Prasad Yadav governed on a backward-caste coalition that treated both the Congress and the BJP as upper-caste formations to be kept from power. In West Bengal, the Left Front’s machine had held the state since 1977 and answered to no Delhi arithmetic. In Tamil Nadu, the Dravida parties alternated in office and admitted no national party as more than a junior passenger. A party system of this kind does not produce national majorities; it produces state verdicts that must then be aggregated by negotiation in Delhi. The BJP’s 161 seats were an aggregation of state verdicts favourable to it. The missing 111 were state verdicts that had gone to parties whose state-level contests were defined against the BJP or against each other, and no speech, ministry or manoeuvre in May 1996 could renegotiate eighty separate state contests from the floor of the Lok Sabha in thirteen days. The renegotiation required the slower instruments the party then took up: state-level alliances, seat adjustments and a national programme priced for partners.

What majority mark did a government need in the Eleventh Lok Sabha?

A government needed 272 seats in the Eleventh Lok Sabha, one more than half of the 543 elected seats recorded in the Election Commission’s statistical report for 1996. The BJP’s 161 seats left it 111 short of that mark, and the party’s pre-poll alliance, credited with 187 seats in the Commission’s alliance tally, stood 85 seats below the line.

The gap of 111 seats is the single number that governs the thirteen days. It was too large for floor management, too large for the inducements a new government can credibly offer, and far too large to close with the small parties and independents who usually settle such questions at the margin. Vajpayee and the BJP leadership, L.K. Advani as party president above all, understood the number before the oath was taken. The decision to accept the invitation anyway, and the conduct of the government during its thirteen days, only make sense as a deliberate positioning exercise by a party playing for the next parliament rather than the present one. The sections on the invitation and on the resignation speech take up that deliberateness in detail.

Two Precedents: 1989 and 1991 Taught Every Party the Minority Game

No participant in May 1996 was improvising. The hung house had been rehearsed twice in the preceding seven years, and both rehearsals taught lessons that the actors of 1996 applied consciously. In December 1989, V.P. Singh formed the National Front government with 143 Janata Dal seats, sustained from outside by the BJP with 85 seats and by the Left parties with their combined bloc, two backers who agreed on nothing except the exclusion of the Congress. That arrangement governed for eleven months, passed a budget, implemented the Mandal recommendations and collapsed in November 1990 when the BJP withdrew its backing over the Ram Janmabhoomi campaign and the arrest of L.K. Advani in Bihar. The 1989 precedent established that a minority government in the post-Congress-system house was a workable instrument, provided an outside backer existed whose interests the government could serve without absorbing it.

The 1991 precedent was stronger still. P.V. Narasimha Rao formed a Congress government in June 1991 with 232 party seats after the election that followed Rajiv Gandhi’s assassination, governed as a minority for long stretches, survived a no-confidence motion in July 1993 by fourteen votes, and completed the full five-year term in May 1996. The 1991 minority-government precedent demonstrated that a determined minority government, working the arithmetic session by session and issue by issue, could outlast a parliament. Rao’s achievement made every party in 1996 fluent in minority mechanics: outside backing, issue-based arrangements, the careful courting of regional parties whose demands were material rather than ideological. It also raised the standard against which the BJP’s thirteen days would be judged. A minority government was no longer an anomaly to be pitied. It was a craft to be practised, and the BJP in May 1996 visibly lacked the craft’s first requirement, a backer.

The deeper architecture within which these experiments ran had been set after 1984. The post-Indira institutional settlement left behind a polity in which the presidency’s discretion in a hung house, the Speaker’s control of floor procedure and the courts’ supervision of defections all carried weight that the era of manufactured majorities had kept dormant. By 1996 those institutions were active players in government formation. President Sharma’s invitation, its conditions and its deadline structured the thirteen days as firmly as any party decision did, and the anti-defection provisions of the Tenth Schedule, in force since 1985, closed the older road to a majority through the purchase of individual members. The BJP’s problem in 1996 was therefore institutional before it was numerical: the system offered exactly two legitimate roads to 272, a pre-poll alliance or a post-poll coalition negotiated between parties, and the BJP had travelled neither far enough.

Government Formation as Constitutional Practice

The discretion President Sharma exercised in May 1996 had been theorised before it was tested. The Sarkaria Commission on Centre-State relations, reporting in 1988, had addressed the cognate problem of the governor in a hung assembly and recommended an order of preference for invitations: a pre-poll alliance commanding a majority first, then the single largest party with the backing of others, then combinations able to demonstrate their numbers, with the floor of the House as the final and only legitimate test. The anti-defection law of 1985 had meanwhile outlawed the older technology of majority-making, the retail purchase of members, so that combinations had to be negotiated wholesale between parties or not at all. President R. Venkataraman’s conduct in 1989 supplied the operative precedent at the Union level: the Congress, as the largest party, was invited first and declined, and V.P. Singh’s National Front was then commissioned on the strength of its demonstrated outside backing. By 1996 the elements of a constitutional practice for hung houses were therefore settled: invitation by size, a deadline for proof, the floor test as arbiter, and no majority recognised that had not been shown in the House. The practice mattered to the thirteen days because it made the BJP’s failure a public and certified event, conducted under rules every party accepted, and therefore made the party’s conduct under failure a public and certified credential. A party that lost quietly under those rules had passed an examination its rivals had set and marked. A party that lost amid disorder or attempted purchase would have failed it for a generation. The constitutional practice of the hung house is the frame that converted Vajpayee’s thirteen days from an embarrassment into an asset, and the BJP’s leadership showed, in its acceptance of the invitation’s terms and its refusal to evade the deadline, that it understood the frame precisely.

The Invitation: President Sharma and the Largest-Party Norm

President Shankar Dayal Sharma’s handling of the 1996 verdict followed the convention that had been consolidating since 1989: where no party or pre-poll alliance commanded a majority, the leader of the single largest party received the first invitation and a stated period within which to demonstrate majority backing on the floor. The results were complete by the middle of May 1996. The United Front constituents, the Congress and the Left were already in negotiation with one another, but none had assembled a claim, and the BJP’s 161 seats made Vajpayee’s claim the only one that could be formalised. The invitation issued on 15 May 1996, the oath followed at Rashtrapati Bhavan on 16 May 1996, and the President set the government a deadline of 31 May 1996 to prove its majority in the Lok Sabha. Vajpayee, returned to the Lok Sabha from Lucknow, took office at seventy-one as the first prime minister drawn from a party of the Hindu nationalist tradition, a fact whose symbolic charge every participant registered and no participant mentioned in the formal exchanges.

When was Atal Bihari Vajpayee first sworn in as Prime Minister?

Atal Bihari Vajpayee was first sworn in as Prime Minister on 16 May 1996, after President Shankar Dayal Sharma invited the leader of the single largest party in the hung Eleventh Lok Sabha to form a government. The President required the government to prove its majority on the floor of the House by 31 May 1996.

The invitation deserves reading as a constitutional act rather than a courtesy. President Sharma, a former Congress vice-president governing his discretion under the scrutiny of every party, conferred on the BJP something the party had never possessed: the status of a government-in-waiting recognised by the head of state, the right to negotiate from inside office, to summon the machinery of the Union and to be treated by foreign governments and by the domestic press as the incumbent power. The thirteen days spent that recognition carefully. The BJP leadership calculated, correctly as 1998 would show, that the legitimacy conferred by the invitation was a wasting asset only if the government disgraced itself in office, by clinging, by purchasing members or by collapsing in disorder. A government that behaved with restraint, argued its case in the House and surrendered office on schedule would bank the legitimacy and carry it into opposition. The invitation, in short, handed the BJP an examination in governability, set by the President and marked by the regional parties. The party’s conduct in office was its answer paper.

The Blocs Counted: 187, 192 and 140

The House that President Sharma surveyed in May 1996 can be counted in blocs, and the bloc count shows why the invitation’s outcome was overdetermined. The BJP and the parties allied with it before the election, the Shiv Sena, the Samata Party, the Shiromani Akali Dal and the Haryana Vikas Party among them, commanded 187 members in a House of 545. The thirteen parties that constituted the United Front in the days after the verdict commanded 192 among them: the Janata Dal’s 46, the CPI(M)’s 32, the Tamil Maanila Congress’s 20, the Dravida Munnetra Kazhagam’s 17, the Samajwadi Party’s 17, the Telugu Desam’s 16, the Communist Party of India’s 12, and the smaller Left and regional contingents behind them. The Congress and its associates held 140. The residue, some two dozen members of small parties and independents, could ornament a majority but could not create one. The United Front’s 192, added to the Congress’s 140 in outside backing, produced the 332 members the Front could parade as its demonstrated strength, a number comfortably past the mark and the basis on which Deve Gowda’s invitation followed Vajpayee’s resignation as a matter of arithmetic course.

Set against that count, the BJP’s position was not a near miss. The government began 85 seats short with its own alliance complete, and every road upward ran through a member of the 192 or through the Congress itself. The Front’s members had organised their combination in explicit opposition to the BJP’s inclusion; the Congress had resolved to keep the BJP from power as the one proposition on which its factions agreed. For the government to survive its confidence motion, it needed not a handful of strays but the defection of an entire organised bloc from a combination built to exclude it, in a House governed by an anti-defection law that penalised exactly such movement by members acting alone and rewarded only party-level decisions. Party-level decisions were available in principle: a party could merge, split by the statutory fraction or change its allegiance as a party. The thirteen days were consumed in discovering whether any party would so decide, and the answer, delivered well before the debate, was that none would. The confidence debate of 27 and 28 May was therefore a ratification ceremony for a count that had been completed in the preceding week, and Vajpayee’s speech was delivered by a man who had read the count and chosen to address the audience beyond it.

Why the BJP Accepted a Doomed Invitation

A question the standard telling seldom asks is why the BJP accepted the invitation at all. The arithmetic was public, the refusals were predictable, and a party jealous of its dignity might have declined the commission, left the field to the United Front’s negotiations and conserved its momentum for the next election, which the Front’s fragility promised within a few years. Biographical accounts of Vajpayee and histories of the party record that the question was argued inside the BJP in the days after the results, and that the argument for acceptance, carried by Vajpayee and Advani together, rested on three calculations. The first was constitutional propriety as strategy: the largest party that refuses the President’s invitation surrenders the claim to be a party of government, and the refusal would be quoted against the BJP at every future formation. The second was intelligence: thirteen days in office would reveal, on the record and at no cost in concessions, exactly which parties would combine with the BJP, on what stated grounds they refused, and which of the grounds were negotiable; opposition would have left the party guessing at the answers it obtained in May 1996 for free. The third was the demonstration effect on the party’s own electorate and on the uncommitted voter: a BJP government that behaved constitutionally under a deadline was the single most efficient refutation of the charge, fatal since 1992, that the party could not be trusted with the state. Each calculation was vindicated by the event. The refusals were catalogued and priced; the conduct was certified by opponents; and the electorate of 1998 returned the party with twenty-one more seats and a queue of partners. Acceptance of a doomed invitation was the most consequential decision of the thirteen days, and it was taken before the government existed, which is why the government’s conduct in office followed a script of restraint that surprised only those who had mistaken the acceptance for optimism.

Thirteen Days in Office: A Ministry Without a Coalition

The ministry Vajpayee assembled on 16 May 1996 was built for a government that expected to expand. Senior figures of the party took the principal portfolios, expansion was held in reserve against the successful confidence vote, and the government settled into a discipline that its critics found disarming and its backers found frustrating: no major policy initiative was announced, no provocative appointment was made, and no attempt was mounted to split any party whose members might be detached. The restraint was partly arithmetic honesty. A council of ministers that might cease to exist within a fortnight had no mandate for irreversible decisions, and the constitutional conventions governing caretaker conduct, though not yet formally in force, were observed in anticipation. The restraint was also strategy. Every day of quiet incumbency contradicted the caricature, standard in opposition rhetoric since the demolition of the Babri Masjid on 6 December 1992, of a BJP government as an imminent majoritarian emergency.

The government’s real work in those thirteen days was the search for numbers, and the search failed in full view. Emissaries of the BJP sounded the Telugu Desam Party of N. Chandrababu Naidu, which held 16 seats; the Tamil Maanila Congress of G.K. Moopanar, which held 20; the Asom Gana Parishad of Prafulla Kumar Mahanta, which held 5; and smaller formations whose adhesion would have signalled momentum. The arithmetic of even total success was unforgiving: the BJP’s 161 and its pre-poll allies’ tally together stood near 187, the three courted parties could add only 41, and the residue still required parties that were organising the alternative government. But the refusals were not primarily about arithmetic. Each courted party faced its own electorate. The Telugu Desam had built its Andhra Pradesh politics on a secular self-image and a direct rivalry with the Congress that the BJP could not mediate. The Tamil Maanila Congress had broken from the Congress on the eve of the election over the Congress’s alliance with the AIADMK in Tamil Nadu, and its identity rested on a Dravidian, non-Congress, non-BJP claim to Tamil pride that adhesion to a BJP government would have dissolved. The Asom Gana Parishad guarded an Assamese regionalism wary of every Delhi formation. Behind each refusal stood 6 December 1992. The demolition at Ayodhya, and the communal violence that followed it across the country, had fixed the BJP in the calculations of regional parties as a partner whose embrace cost more votes than it delivered. The thirteen-day government was the BJP’s discovery, conducted at the summit of the state, that the party’s road to durable power ran through a renegotiation of that price.

The Federal Proof of Concept: The BJP in the States

The party that could not assemble a coalition in Delhi in May 1996 was already governing in coalition in the states, and the state record is part of the explanation for the speed of the national adaptation. In Maharashtra, the BJP had shared power with the Shiv Sena since March 1995, accepting the junior partner’s portion in a government headed by the Sena’s Manohar Joshi, an arrangement that required the party to govern alongside a regional formation whose style and programme diverged from its own at nearly every point and that survived its full term. In Delhi, the BJP governed the state through the hawala season’s casualties, Madan Lal Khurana’s resignation as chief minister in the wake of the Jain diary allegations and Sahib Singh Verma’s succession in February 1996. In Rajasthan, Bhairon Singh Shekhawat’s government rested on the party’s accommodation with the state’s regional and caste notables. In Gujarat, by contrast, the party’s state unit tore itself apart in September 1996, when Shankersinh Vaghela broke away to form a government with Congress backing, a demonstration from within the family that organisational indiscipline could cost the BJP a state it had won outright. The pattern across the four theatres taught the same lesson the Lok Sabha taught in May: where the BJP practised partnership, accommodation and discipline, it governed; where it relied on its own weight and its own factions, it bled. The national leadership’s post-1996 strategy was the state experience generalised, and the partners who negotiated with the BJP in 1997 and 1998 could inspect the Maharashtra record, five years of shared government with a difficult ally, as the party’s bond. Coalition credibility, like the personal credibility Vajpayee carried, was an asset accumulated in the states before it was spent in Delhi.

A Party Between Movement and Government

The isolation of May 1996 was not only a judgment other parties passed on the BJP. It was also the visible form of a tension inside the Hindu nationalist movement between its character as an ideological movement, with a programme held in trust by the Rashtriya Swayamsevak Sangh and a cadre schooled to regard compromise as dilution, and its character as a party of government, for which the programme was an opening bid in negotiations the movement did not control. The BJP that contested the 1996 election carried the full inheritance of the Ayodhya mobilisation: the rath yatra of 1990, the demolition of 1992, the dismissal of its four state governments by the Rao ministry in the demolition’s aftermath and the party’s subsequent return in state after state, Madhya Pradesh and Rajasthan and Himachal Pradesh among them, on the strength of the same mobilisation. To its cadres, the road to Delhi ran through the intensification of that politics. To the arithmetic of the Eleventh Lok Sabha, the road ran through its moderation. The thirteen days forced the choice into the open at the highest level of the movement, with the prime ministership itself as the stake.

Vajpayee’s position inside that tension was the party’s bridge across it. He had been a founding figure of the Bharatiya Jana Sangh in 1951 and of the BJP in 1980, a pracharak-trained organiser by formation and a parliamentarian by vocation, and his authority with the cadre rested on a record no rival inside the party could match; his moderation was therefore legible to the movement as strategy rather than apostasy, a distinction on which the entire adaptation of 1996 to 1998 depended. Advani, the movement’s organiser and the architect of its mobilisation decade, supplied the complementary guarantee from the other side: when Advani as party president argued that the BJP had reached the ceiling of its own strength and must learn coalition, the argument carried the weight of the man who had built the strength being declared insufficient. The dual leadership that had powered the party’s rise, Vajpayee’s face and Advani’s machine, converted itself after May 1996 into the dual instrument of its transformation, the face to reassure partners and the machine to deliver the organisation’s consent. Movements rarely survive the discovery that their founding programme cannot win power unamended. The BJP’s survival of that discovery, transacted between 1996 and 1998 under the pressure of the thirteen days, is the organisational fact beneath the parliamentary story.

The Isolation: Untouchability and the Organised Alternative

Christophe Jaffrelot, in The Hindu Nationalist Movement and Indian Politics, published in the election year itself, 1996, supplied the analytic term for the BJP’s condition: the party was integrated into the party system as a competitor and excluded from it as a partner, a position Indian political commentary of the period named without embarrassment as political untouchability. Jaffrelot’s study traced the Hindu nationalist movement’s alternation between mobilisation and moderation across seven decades and found the BJP of the mid-1990s suspended between the two: too large to ignore, too marked by the Ayodhya campaign to embrace. The refusals of May 1996 were that suspension rendered in parliamentary arithmetic. Parties that might gladly have used the BJP’s 161 seats to exclude the Congress discovered, at the moment of decision, that association carried a price with their own voters, their own cadres and their own coalition partners that no portfolio could repay.

The alternative government was not waiting passively for the BJP to fail; it was built in parallel, and its construction is part of the thirteen days’ record. The Janata Dal, the Left parties, the Telugu Desam, the Tamil Maanila Congress, the Dravida Munnetra Kazhagam, the Samajwadi Party and their associates, thirteen parties in the eventual count, constituted themselves as the United Front during the very days the BJP government sat in office, with Chandrababu Naidu as the Front’s convener and a steering committee preparing a common minimum programme. The Congress, holding 140 seats and the memory of its July 1993 survival by fourteen votes, declined to head a government it could not sustain and resolved instead to back the United Front from outside, a decision the party’s leadership under P.V. Narasimha Rao reached as the confidence debate approached. H.D. Deve Gowda, the Janata Dal chief minister of Karnataka, emerged as the Front’s candidate for prime minister after V.P. Singh and the CPI(M) leader Jyoti Basu each declined the office. Every element of this architecture, the Front, the outside Congress backing, the agreed candidate, was in place before Vajpayee rose to reply to the confidence debate on 28 May 1996. The BJP government did not fall for want of effort or want of days. It fell because the rest of the party system had organised its exclusion faster than it could organise its own inclusion, and the lesson the party drew from the speed of that organisation shaped the next twenty-two months.

The Confidence Debate of 27 and 28 May 1996

The government’s confidence motion came before the Lok Sabha on 27 May 1996, inside the deadline President Sharma had set, and the debate ran across two days. A CRS-based study of confidence motions published by PRS Legislative Research records the essential procedural fact: the 1996 motion was discussed for some eleven hours and was never put to a vote, because the prime minister announced his resignation from the floor of the House at the debate’s close. The debate’s form mattered to its substance. A confidence debate in a hung house is not an inquiry into a government’s record, for a thirteen-day government has none. It is an auction of declarations, in which each party states, on the permanent record of the House, the terms on which it will and will not combine. Speaker after speaker from the Congress, the Janata Dal, the Left and the regional parties declared against the government, and the declarations were directed as much at one another, and at the electorate listening beyond the chamber, as at the treasury benches.

Who presided over the Eleventh Lok Sabha during the confidence debate?

P.A. Sangma of the Congress presided over the Eleventh Lok Sabha as Speaker during the confidence debate of 27 and 28 May 1996, having taken the chair on 23 May 1996. His election as Speaker, secured with broad backing across the House, was itself an early sign that the non-BJP parties could combine when a prize was within reach.

The opposition’s argument, rendered at its strongest, rested on two legs. The first was arithmetic legitimacy: a government commanding 161 of 543 seats, opposed by the declared holders of nearly every other seat in the House, stood the presumption of parliamentary government on its head, and the largest-party convention that had seated it could not oblige any member to sustain it. The second was the character of the BJP itself. Opposition speakers, Somnath Chatterjee of the CPI(M) prominent among them in the debate, held that the party’s record on Ayodhya disqualified it from the trust of a plural polity, and that the moderation of its prime ministerial face did not alter the commitments of the movement behind him, a charge the Sangh’s critics had pressed since 1993 and would press again in 1998. The treasury benches answered with the counter-charge that a combination assembled from thirteen parties and an outside backer, united only in exclusion, had no mandate to govern either, and that the House was witnessing not the defence of secularism but the arithmetic of a cartel. Both arguments were honestly held and both were, in their way, correct, which is what made the debate an accurate map of the party system rather than a mere formality. Vajpayee’s reply closed the debate on 28 May 1996, and the reply is the document on which the structural reading of the thirteen days stands or falls.

The Resignation Speech: A Coalition Doctrine Spoken from Defeat

The reply Vajpayee delivered to the Lok Sabha on 28 May 1996, printed in the Lok Sabha debates of that date, ran for about an hour and performed three distinct operations, which the coalition-strategy timeline presented later in this article sets out in sequence. The first operation was refusal. Vajpayee announced that the government would resign rather than purchase a majority, and he framed the refusal as doctrine rather than disappointment: the methods by which 111 missing members might have been detached from their parties were available in the market of the House, and the BJP declined to trade in that market. Contemporaries across the divide registered the refusal; the press coverage of 29 May 1996 recorded opposition members conceding that no attempt at horse-trading had stained the thirteen days. The second operation was prophecy. Turning to the combination preparing to succeed him, Vajpayee warned the House that a government assembled to exclude the BJP carried its instability inside itself: governments would come and go, he argued, but a coalition whose members had to run to the Congress for every decision had mistaken exclusion for a programme. Reports of the speech preserve the sentence in near-identical form across publications: if a government was to be formed by leaving the BJP out, he saw no sign of its stability; the birth would be difficult, and survival after the birth more difficult still. The third operation was the famous peroration, delivered in Hindi and quoted in every account since: Sarkarein aayengi, jayengi; partiyan banengi, bigdengi; magar yeh desh rehna chahiye, iska loktantra amar rehna chahiye. Governments will come and go, parties will be formed and broken, but this country must endure, and its democracy must remain immortal.

Read as rhetoric, the speech is a masterpiece of the genre, and the standard treatment reads it as rhetoric only, the graceful exit of a parliamentarian of the old school. Read as a strategic document, which is the reading this article defends, the speech is a set of three coalition principles stated on the record of the House and addressed, over the heads of the members present, to the regional parties who would hold the balance in every future parliament. The first principle concerned legitimacy: the BJP accepted, in terms, that the autonomy of the regional parties was a permanent feature of the polity and not an aberration to be waited out, and that a party seeking their partnership would have to earn it by conduct, beginning with a refusal to subvert their members. The second principle concerned programme: Vajpayee’s account of his own government’s restraint, and his pledge that the BJP in opposition would back measures it judged to serve the country whoever moved them, sketched the working rule of the later common programme, that a coalition governs on what its members agree and shelves what they do not. The third principle concerned arithmetic itself: the speech insisted that numbers assembled without agreement were not a majority in any sense that could govern, so that the road to power for the BJP ran through negotiated agreement before the election, not through improvisation after it. A party that had just failed to form a durable government thus announced, at the moment of failure, the three rules under which it would succeed in forming one twenty-two months later. That is the sense in which the resignation speech was the foundation document of the 1998 National Democratic Alliance, and it is why the speech, rather than any act of the thirteen-day ministry, carries the article’s claim.

The context of delivery sharpens the point. Vajpayee remained prime minister in a caretaker capacity after tendering his resignation to President Sharma on the evening of 28 May 1996, and his formal tenure is accordingly listed in reference works as running to 1 June 1996, the day Deve Gowda took oath. The speech belongs to 28 May, to the floor of the House and to the debate it closed. Its author knew, as he spoke, that the audience that mattered was not voting that evening. The speech was an application for a future hearing, filed in the only registry whose records no later negotiation could amend, the debates of the Lok Sabha.

The speech’s textual afterlife reinforced its function as a charter. It was printed in full in the Lok Sabha debates, reproduced by the BJP in its documentation of the period, and anthologised in the collected volumes of Vajpayee’s speeches through which the party’s own cadres studied it. A doctrine that exists as a citable text behaves differently from a doctrine remembered as an attitude: it can be invoked by a negotiator against the party that proclaimed it, which is precisely the use the partners of 1997 and 1998 made of the speech’s pledges, and it disciplines the proclaiming party’s own future conduct, since a public text cannot be quietly revised. The BJP’s negotiators in the winter of 1997 and 1998 therefore worked inside constraints their leader had published in May 1996, on programme, on the treatment of partners and on the refusal of retail defection, and the partners’ knowledge that the constraints were public raised the value of every assurance the negotiators gave. The speech’s preservation as text is a small fact with large consequences, and it completes the explanation of why this particular resignation, among the several in the republic’s history, became a foundation document rather than a farewell address.

The Speech in the Country: Reception and Afterlife

The speech’s first audience was the House; its decisive audience was outside it. Parliamentary proceedings reached the country through the press of 29 May 1996 and through the broadcast record, and the coverage fixed the episode’s public meaning within a day: a prime minister who had resigned rather than bargain, an opposition that had won the arithmetic and been denied the moral of the victory, and a standard of conduct in defeat against which the succeeding government’s manoeuvres would be measured. Editorial judgment across papers that had opposed the BJP’s politics conceded the manner of the exit, and the concession travelled. Through the United Front’s twenty-one months of withdrawals and ultimatums, the comparison wrote itself in the public mind: the government that had lasted thirteen days had behaved with more institutional decorum than the combinations that lasted ten months at a stretch. Political capital of that kind is difficult to manufacture and impossible to purchase, and the BJP drew on it openly in the campaign of 1998, in which Vajpayee’s portrait, rather than the party’s symbol, carried the alliance’s appeal in constituency after constituency. The party’s campaign literature of 1998 returned repeatedly to the resignation as the credential of its leader’s fitness to head a coalition: the man who would not buy a majority in 1996 could be trusted to hold one together in 1998, because his partners had watched him decline the instruments with which majorities are usually stolen. The speech thus acquired the afterlife that distinguishes a text from a performance. It was quoted in the alliance’s negotiations as a bond, cited by partners as the undertaking on which they had relied, and preserved in the Lok Sabha’s record as the doctrine’s charter, available to any reader who wished to test the government of 1998 against the promises of the defeat of 1996.

Four Days of Interregnum and the United Front Answer

The interval between the resignation of 28 May 1996 and the oath of 1 June 1996 displayed the rival coalition logic in working order. The United Front’s steering committee completed its arrangements, the Congress formalised its outside backing, and President Sharma invited Deve Gowda, who resigned the chief ministership of Karnataka and was sworn in as prime minister on 1 June 1996 with eleven days to prove his majority, which the Front’s numbers, reinforced by the Congress’s 140, comfortably supplied. The United Front government that followed was in one respect the anti-BJP’s answer to the question the thirteen days had posed: it demonstrated that the post-Mandal arithmetic could also produce a governing bloc of the Janata parivar, the Left and the regional parties, held in office by a Congress that preferred external influence to internal responsibility. In another respect the Front was a confirmation of Vajpayee’s prophecy from within the experiment itself. Its common minimum programme papered over the differences among thirteen parties; its prime minister held office at the pleasure of the Congress’s outside backing; and when the Congress withdrew that backing in the spring of 1997, Deve Gowda fell on 21 April 1997, to be replaced by I.K. Gujral of the same Front, whose government the Congress in turn brought down in November 1997, forcing the election of February and March 1998.

How long did the Deve Gowda government last?

The Deve Gowda government lasted from 1 June 1996 to 21 April 1997, a little under eleven months, before the Congress withdrew the outside backing on which the United Front depended. I.K. Gujral succeeded him the same day and held office until 19 March 1998, when the next Vajpayee government took oath.

The comparison that the Eleventh Lok Sabha invites is with the earlier experiment in anti-Congress combination government. The 1977 Janata coalition’s fragility had dissolved from within, as the merged party’s constituent ambitions consumed the government that housed them; the United Front dissolved from without, as its external backer’s calculations changed. The mechanism differed and the lesson compounded: a combination whose members are united by what they exclude must either convert exclusion into programme at speed or be consumed by the arithmetic that assembled it. The United Front never made the conversion, because its members’ programmes contradicted one another on the questions, economic liberalisation above all, that government forced upon them. The BJP, in opposition through 1996 and 1997, watched the demonstration and drew the operational conclusion that its own combination, when it came, would have to be assembled before the election, bound by a programme negotiated in advance, and led by a figure whose personal credibility could hold partners whose ideologies diverged. The thirteen days had stated the rules. The United Front’s twenty-one months supplied the worked example of their violation.

The Other Side’s Seminar: Basu and the Historic Blunder

The United Front’s construction contained one refusal as consequential as any the BJP received. When the Front’s constituents settled their candidate for prime minister in the last days of May 1996, the name with the widest acceptance was Jyoti Basu, the CPI(M) chief minister of West Bengal since 1977 and the Left’s most authoritative figure. Basu was willing; his party was not. The CPI(M)’s central committee voted against permitting its leader to head a government the party could not control, preferring the doctrinal safety of outside influence within the Front to the responsibility of leading it, and the Front turned to Deve Gowda. Basu’s later verdict on his party’s decision entered the language: he called it a historic blunder, a judgment he repeated in interviews for the rest of his life. The episode matters to the thirteen days because it completes the symmetry of May 1996. Every formation in the House was conducting the same seminar in coalition logic under the same arithmetic, and every formation was learning at its own speed. The BJP learned that partnership must be negotiated and priced. The Congress learned that outside backing is influence on loan. The CPI(M) learned, by Basu’s own testimony, that the refusal of power at the moment it is offered is also a decision with a price, payable in the currency of relevance. The party system that emerged from the Eleventh Lok Sabha was educated less by its victories than by the specific refusals of that single month, which is one more reason the month repays the attention the footnote treatment denies it.

The Congress Decision: Power Without Office

One actor’s conduct across the thirteen days rewards separate attention, because the Congress wrote the rules of the interregnum as surely as the BJP wrote the doctrine that outlasted it. The Congress of May 1996 held 140 seats, the second-largest bloc in the House, and a plausible claim to attempt a government of its own at the head of the United Front constellation. The claim was never pressed. P.V. Narasimha Rao, prime minister until 16 May 1996 and president of his party, led a leadership that read the arithmetic with cold accuracy: a Congress-led minority would depend on the same regional parties that were organising the United Front, would carry the odium of the hawala season into office, and would expose the party to destruction at the floor test or to slow strangulation after it. The party therefore chose the position of outside backer to a government it did not join, influence without office, the power to sustain and to withdraw without the responsibility of governing. The choice was rational in the short account and ruinous in the long one. It taught the regional parties that the Congress could be used as a ladder and discarded, it taught the BJP that the Congress’s fear of office was an asset to be banked, and it began the sequence, the withdrawal from Deve Gowda in April 1997, the destruction of Gujral that November, the defeat of 1998, that ended with the party in opposition for a decade and its leadership passing, in March 1998, to Sonia Gandhi. The thirteen days thus hold a mirror to both national parties: the one that took office without numbers and converted the failure into a doctrine, and the one that declined office with numbers enough to matter and converted the prudence into a decline. Government formation in a hung house rewards the party that treats the event as an investment. In May 1996 only one national party did.

The Strategic Adaptation: June 1996 to March 1998

The twenty-two months between the resignation and the return are the least documented and most consequential stretch of the story. The public record holds the landmarks: the proceedings of the BJP’s National Executive and National Council across 1996 and 1997, preserved in the party’s archives in Delhi, in which the isolation of May 1996 was named as the problem the organisation existed to solve, and the resolutions through which the party repositioned itself from a movement awaiting a majority into a coalition entrepreneur assembling one. Walter Andersen and Shridhar Damle, in The RSS: A View to the Inside (2018), document the Sangh’s side of the recalibration: the Rashtriya Swayamsevak Sangh’s leadership accepted, not without friction recorded in their account, that the road to a BJP-led government ran through alliances with parties indifferent or hostile to the Sangh’s cultural programme, and that the governmental wing of the movement would set the terms of that courtship. L.K. Advani, party president until 1998, carried the organisational argument that the BJP’s growth had reached the ceiling of its own mobilisation, and that the next increment of power would be rented from partners or not obtained at all.

The adaptation proceeded on three tracks, each traceable to a principle of the resignation speech. The first track was the conversion of refusals into relationships. In Punjab, the Shiromani Akali Dal of Parkash Singh Badal, whose eight members had sat with the BJP’s pre-poll tally in 1996 without rescuing it, deepened the partnership into a state alliance that swept the Punjab assembly election of February 1997 and installed Badal as chief minister, giving the BJP a working model of shared government with a regional party whose base, Sikh farmers of Punjab, owed nothing to Hindu nationalist mobilisation. In Maharashtra, the standing partnership with the Shiv Sena, in government in the state since 1995, supplied a second working model. In Orissa, the Biju Janata Dal, founded by Naveen Patnaik in December 1997 from the wreckage of the state Janata Dal, entered the BJP’s orbit for the 1998 election rather than the United Front’s. In Andhra Pradesh, Chandrababu Naidu, the convener of the United Front itself, watched his Front disintegrate through 1997 and began the recalibration that would carry the Telugu Desam into outside backing of the BJP-led government in 1998. Each of these movements was a transaction with its own local logic. Together they describe a party system learning, party by party, that the untouchability of 1996 had been a price quoted in a particular market, the immediate aftermath of Ayodhya and a hung house, and that the price moved as the market changed.

The second track was programme. The BJP’s negotiators, working through the winter of 1997 and 1998, drafted the document that became the National Agenda for Governance, the common programme of the alliance that contested the 1998 election. Its method was the second principle of the resignation speech reduced to a drafting rule: the alliance would govern on the ground its members shared, economic reform, federal courtesies, administrative repair, and the BJP’s three signature commitments that its partners could not accept, the construction of a Ram temple at Ayodhya by legislation, the abrogation of Article 370 and a uniform civil code, were placed outside the common agenda for the life of the coalition. The third track was the candidate himself. The alliance of 1998 was assembled explicitly around Vajpayee, whose thirteen days had converted him from the BJP’s most acceptable face into a statesman whose personal assurances partners could bank, and whose name headed the alliance’s appeal in states where the BJP’s organisation barely existed. The adaptation was thus not a softening improvised for an election. It was the institutionalisation, across organisation, programme and leadership, of a doctrine that had been announced from the floor of the Lok Sabha on 28 May 1996.

The adaptation also had a discursive track, conducted in the party’s public argument with itself. Through 1997 the BJP’s senior figures used the party’s council sessions and the press to normalise a proposition that would have been heresy to the cadre of 1992: that coalition government was not a regrettable interlude between majorities but the normal form of Indian national government for the era, and that a party aspiring to lead it had to present itself as a coalition-maker before it presented itself as a claimant. The proposition’s acceptance can be tracked in the alliance’s own campaign of 1998, fought not on the BJP’s symbol alone but on Vajpayee’s name and the National Agenda’s text, with allies contesting on their own symbols under seat adjustments negotiated state by state. The machinery of seat adjustment, routine in state politics and resisted for decades in national contests by parties dreaming of single-party majorities, was itself an admission written into practice: the BJP of 1998 conceded constituencies it might have contested, husbanded its strength where it was strong and borrowed its partners’ strength where it was weak. The thirteen days had announced that arithmetic alone could not govern India. The 1998 campaign was the first the BJP fought as if it believed its own announcement.

The Regional Ledger: Prices Quoted in 1996, Prices Paid in 1998

The adaptation can be audited party by party, and the audit shows a market repricing rather than a conversion. The Telugu Desam Party, whose 16 seats Naidu withheld in May 1996 while he convened the United Front, moved by stages as the Front decayed: convener of the anti-BJP combination in 1996, neutral arbiter as the Gujral government fell, and outside backer of the BJP-led government in 1998, at a price that respected its Andhra Pradesh rivalry with the Congress and its secular self-description, no ministers in the Union cabinet, the Speakership for its nominee G.M.C. Balayogi, and consultation as of right. The Dravida Munnetra Kazhagam, a pillar of the United Front in 1996, crossed to the BJP-led alliance for the election of 1999 after the Front’s collapse and the AIADMK’s occupation of the 1998 berth in Tamil Nadu, a movement dictated by state rivalry rather than national doctrine and accepted by the BJP on exactly those terms. The Tamil Maanila Congress, whose 20 seats had been the largest single refusal of May 1996, never entered the BJP’s alliance; Moopanar’s party kept its distance through 1998 and faded with its founder’s death in 2001, a reminder that the courtship of 1996 to 1998 succeeded selectively, where state arithmetic rewarded it, and failed where it did not. The Asom Gana Parishad likewise held apart in 1998. The ledger’s pattern is the doctrine’s vindication in transactional form: the BJP stopped asking regional parties to endorse its programme and started asking what each party’s state contest required, then paid in the currencies the answers specified, portfolios, speakerships, seat concessions, programmatic silence. Untouchability had been the price of endorsement demanded in 1996. Partnership in 1998 was priced in interests, and interests can be met.

The Coalition-Strategy Timeline, 1996 to 1998

The argument of this article can be stated as a sequence, and the sequence can be checked against the dated record. Researchers reconstructing these twenty-two months work through the Election Commission’s statistical report on the 1996 general election, the Lok Sabha debates of 27 and 28 May 1996, the BJP’s published resolutions of 1996 and 1997 and the text of the National Agenda for Governance, much of that material circulating in scanned compilations that readers open in tools such as an office file viewer before they cite it. The table below, the coalition-strategy timeline from May 1996 to March 1998, sets out the phases of the adaptation with the named decision and the named outcome that define each phase, so that the structural reading can be cited against the record rather than asserted over it.

Phase	Dates	Named decision	Named outcome
The claim	15 and 16 May 1996	President Sharma invites Vajpayee; the oath is taken	A BJP government holds office with 161 party seats, 111 short of the majority mark
The refusal	16 to 28 May 1996	The Telugu Desam, the Tamil Maanila Congress and the Asom Gana Parishad decline to join	The coalition arithmetic freezes below 200 and the alternative organises
The doctrine	28 May 1996	Vajpayee closes the confidence debate and tenders his resignation	The BJP exits without a vote and without horse-trading, its three coalition principles on the record
The reassessment	June 1996 to 1997	BJP National Executive and National Council proceedings address the isolation	Alliance-building replaces majority-waiting as the party’s stated strategy
The courtship	1997 to early 1998	Engagement with the Akali Dal, the Biju Janata Dal at its founding and the Telugu Desam under Naidu	Parties that held back in 1996 negotiate entry or outside backing for 1998
The consummation	March 1998	The National Agenda for Governance is adopted and Vajpayee takes oath on 19 March 1998	A coalition of more than twenty parties takes office under the 1996 doctrine

The timeline’s evidentiary value lies in the direction of causation it makes visible. Each phase answers the phase before it: the refusal produced the doctrine, the doctrine governed the reassessment, the reassessment organised the courtship, and the courtship supplied the consummation. A footnote does not behave in that way. An accident of a hung parliament explains the claim and the refusal; it cannot explain why a party emerged from a thirteen-day failure with a doctrine, a drafting process and a courtship list that correspond, item by item, to the obstacles the failure had exposed.

The Winter of 1997: The Doctrine’s Final Examination

Between the doctrine’s announcement and its consummation stood an examination neither party set but both sat: the collapse of the United Front experiment in the winter of 1997 and 1998, conducted in full public view. The sequence is worth setting out with its dates, because the BJP’s negotiators treated each stage as evidence. In April 1997 the Congress withdrew its backing from Deve Gowda over grievances its leadership had accumulated against his government’s style and substance, and the Front replaced its prime minister rather than face the country, installing I.K. Gujral on 21 April 1997 on a renewed Congress undertaking. In November 1997 the Congress withdrew again, this time over the Jain Commission’s interim findings on the Rajiv Gandhi assassination investigation and the Front’s refusal to drop the Dravida Munnetra Kazhagam from the government at the Congress’s demand, and the Gujral ministry resigned on 28 November 1997. The Lok Sabha was dissolved in December 1997 and the election of February and March 1998 followed. Each stage confirmed a clause of the resignation speech. The outside backer behaved precisely as Vajpayee had predicted a backer with influence and no responsibility behaves, sustaining and destroying two governments in nineteen months without once answering to the House for the consequences. The Front’s prime ministers governed at a landlord’s pleasure, which is what the speech’s image of members running to the Congress for every decision had described. And the electorate, presented in 1998 with the bill for the experiment, delivered the BJP its largest tally yet and its partners their entry into office. The BJP’s conduct through that winter was the doctrine applied in opposition: no attempt to topple, no retail bargaining with the Front’s fragments, and a courtship of the future conducted in daylight, through party councils, public resolutions and the drafting of the common programme. By the time the Gujral government fell, the alternative government existed in draft, in personnel and in principle. The examination of 1996 had been set by the President and the arithmetic. The examination of 1997 was set by the United Front, and the BJP passed it by the method the thirteen days had taught: patience as a form of preparation.

March 1998: The Operationalisation

The election of February and March 1998 returned the BJP with 182 seats, still short of a majority, and this time the shortfall did not matter, because the coalition had been built before the counting began. The National Democratic Alliance that took office on 19 March 1998, when Vajpayee was sworn in for the second time, joined more than twenty parties in and around the government: the Shiv Sena, the Samata Party, the Akali Dal, the Biju Janata Dal, the Dravida Munnetra Kazhagam’s rivals of the AIADMK under J. Jayalalithaa, the Trinamool Congress founded by Mamata Banerjee in January 1998, and a tail of smaller formations, with the Telugu Desam Party of Chandrababu Naidu backing the government from outside in exchange for the Speakership of the Lok Sabha for its nominee G.M.C. Balayogi and a recognised voice in the alliance’s counsels. The National Agenda for Governance bound the combination: a negotiated text, adopted by the partners, that committed the government to its shared ground and recorded the exclusion of the three contested commitments from the coalition’s programme. Every feature of the architecture answered a feature of May 1996. Where the thirteen-day government had sought partners after the oath, the 1998 coalition had been negotiated before it. Where the thirteen-day government had faced the House with an arithmetic assembled by invitation, the 1998 government faced it with an arithmetic assembled by agreement. Where Vajpayee in 1996 had been the leader of a party pleading for trust, Vajpayee in 1998 was the named leader of an alliance that had chosen him in writing.

March 1998 also closed the constitutional loop that May 1996 had opened. President K.R. Narayanan, succeeding Sharma in July 1997, conducted the 1998 formation under a refinement of the practice his predecessor had applied: before issuing the invitation, the President required the claimant to produce written evidence of the alliance’s composition and backing, letters from the leaders of the constituent parties recording their agreement to the common programme and to Vajpayee’s leadership. The requirement reads in retrospect like the doctrine of the thirteen days translated into presidential procedure. In 1996 the BJP had been invited on the strength of its size and sent into the House to discover its support; in 1998 it arrived at Rashtrapati Bhavan with its support already negotiated, documented and signed, a coalition on paper before it was a government in office. The distance between the two formations, measured in the twenty-two months between them, is the precise size of the adaptation the resignation speech had prescribed, and the Presidents who presided over the two formations certified both ends of it, the first by setting the examination and the second by recording that the candidate had passed it before the oath was administered.

The 1998 government’s own end, brought down in April 1999 by a single vote after the AIADMK withdrew, and the fuller consolidation that followed the election of October 1999, when the alliance returned with a decisive combined majority and governed to 2004, are treated in later articles of this series. Their relevance here is as a test. The 1999 fall showed that the doctrine of 1996 was a discipline and not a charm: when the programme discipline broke and a partner’s demands exceeded the common agenda, the government fell by one vote, exactly as the arithmetic of coalition required. The 1999 return showed the discipline restored and enlarged. Through both, the operating system installed between May 1996 and March 1998 remained the one the resignation speech had specified, government by negotiated programme under a leader whose credibility was the alliance’s reserve asset.

The Scholarly Disagreement: Footnote or Foundation Document

The disagreement this article adjudicates is between two readings of the same thirteen days. The footnote reading, carried by the standard reference and textbook treatments, holds that the ministry of May 1996 was an interlude produced by a mechanical convention, the largest-party invitation, and terminated by a mechanical fact, the missing 111 seats; on this reading the episode’s only durable product was biographical, a chapter in the making of Vajpayee’s reputation. The structural-signal reading, defended here, holds that the thirteen days were the moment the post-Mandal party system disclosed its operating rules to its largest party, and that the disclosure was acted upon with a fidelity that the 1998 outcome verifies. The scholarly literature, read closely, sides with the second reading on the evidence each camp itself supplies. Jaffrelot’s 1996 study, written before the event, diagnosed the BJP’s exclusion as a structural property of the party system rather than a mood of the moment, which is why the exclusion could only be undone by structural adaptation of the kind the party then undertook. Pradeep Chhibber, in Democracy without Associations (1999), situated the fragmentation of the 1990s in the long decline of party organisation as an intermediary institution, a party system of electoral machines without deep associational roots, in which coalition was not an episode but the resting state of national politics; on Chhibber’s account, the party that learned coalition first would govern the era, and the thirteen days were the BJP’s first lesson. Andersen and Damle, writing in 2018 with access to the Sangh’s internal deliberations, document the adaptation from inside the movement that had to be persuaded to accept it.

There is an analogue inside the earlier history of the Congress system that the structural reading illuminates by resemblance. The 1969 split’s structural and cultural rupture was, in its own news cycle, a party quarrel of a fortnight’s duration, a presidential election, an expulsion, a pair of rival sessions; its consequences reorganised Indian politics for a generation, because the quarrel disclosed and then settled the question of where power inside the ruling party resided. The thirteen days stand in the same relation to the coalition era. Their news value expired on 1 June 1996. Their information content, the price of partnership, the necessity of programme, the insufficiency of arithmetic, governed every government formed for the remainder of the era. The footnote reading mistakes the duration of an event for the measure of its consequences, a mistake the discipline of history exists to correct.

The Complication: Thirteen Days Without a Policy Record

The strongest counter to the structural-signal reading deserves its full force. The government of May 1996 did nothing. It passed no legislation, altered no policy, made no appointment of consequence and left no administrative trace that a later government had to honour or reverse; a reader who measures governments by their outputs will find the thirteen days a blank, and the blank is not a hostile invention but the ministry’s own design. If the episode mattered, the counter runs, it mattered the way a campaign advertisement matters, as positioning in the minds of voters, and positioning is a thinner thing than the structural significance this article claims. The counter is strengthened by a further observation: the adaptation of 1996 to 1998 had many authors and many causes, the United Front’s failures, the Congress’s decline under its post-Rao leadership, the ambitions of regional leaders who found the BJP’s embrace more useful as the Front decayed, and to credit a single speech with the architecture of 1998 is to mistake a symbol for a mechanism.

The answer begins by conceding the premises. Direct policy consequences were nil, and the speech was not the sole cause of the adaptation. The claim survives the concessions because it was never a claim about policy output or sole causation. It is a claim about demonstrated possibility and stated doctrine. The thirteen days demonstrated, to the only audiences whose beliefs governed the future, the other parties and the BJP itself, that a non-Congress, non-Janata bloc could claim the government of India, hold office without disorder and surrender it without disgrace; demonstrations of that kind alter what negotiators in later rounds believe to be available, and beliefs of that kind are the substance of party-system change. The speech then gave the demonstration a text. When the party’s negotiators sat down in 1997 to draft a common agenda and to court the Telugu Desam, they were not improvising a philosophy under pressure; they were executing principles their leader had announced on the record of the Lok Sabha, which partners could quote back to him as security for their own risk. Symbols that can be quoted in a negotiation are mechanisms. The blank policy record, properly read, is part of the evidence: a government that did nothing for thirteen days except demonstrate governability and announce a doctrine had understood, with unusual clarity, what its thirteen days were for.

What the Doctrine Did Not Settle

Honesty about the doctrine’s limits belongs to the adjudication, because the structural reading is weakened, not protected, by claiming more than the record holds. The principles of May 1996 settled how a BJP-led coalition could be formed and held together at the moment of formation. They did not settle how such a coalition would digest the ambitions of partners whose price rose with their indispensability, and the fall of the second Vajpayee government in April 1999, by a single vote, after J. Jayalalithaa’s AIADMK withdrew over demands the common agenda could not accommodate, demonstrated the unsettled residue in the harshest form. Nor did the doctrine settle the internal question it had deferred: the three shelved commitments remained the programme of the movement’s base, and every year of shelving stored a pressure inside the alliance and inside the party that later decades would release in other forms. The doctrine was a technology of coalition, not a resolution of the ideological contest that had made coalition necessary. What the thirteen days settled was narrower and, for the era, sufficient: that the post-Mandal republic would be governed by negotiated combinations, that the BJP intended to be the combination’s centre rather than its excluded pole, and that the terms of its centrality had been stated, on the record, by the man who would embody them. Eras are governed by the settlements they actually reach. The settlement of May 1996 was procedural, and procedure, in a parliamentary republic, is not a small thing. A republic that changes its governing class without a shot, a riot or a constitutional rupture, purely by the working of invitations, deadlines, debates and resignations conducted in public, has demonstrated something about its own durability that longer and louder governments often leave unproven, and the demonstration was among the gifts the thirteen days left to every successor government of the era, including the ones assembled to exclude the party that made it.

Personality as Institution: Why the Doctrine Needed Vajpayee’s Voice

Framework and doctrine aside, the adaptation had a human precondition, and the brief’s secondary framework, personality-as-institution, names it. The coalition logic of 1996 could only be announced credibly by a leader whose personal standing crossed party lines, because the doctrine’s first principle was a promise about future conduct, and promises are believed in proportion to the promiser. Vajpayee’s standing had been accumulating for two decades outside the Hindu nationalist movement’s own constituency: as foreign minister in the Janata government of 1977 to 1979 he had addressed the United Nations General Assembly in Hindi and repaired relations with China and Pakistan to general applause; as leader of the opposition and as the BJP’s parliamentary face through the 1980s and 1990s he had cultivated friendships across the aisle that survived the Ayodhya years; as a poet of published collections he occupied a place in the country’s literary culture that no other party leader of his generation shared. When such a figure announced from the floor of the House that his party would not purchase a majority and would honour the autonomy of the regional parties, the announcement drew on a personal account that the party, as an institution, could not have opened on its own credit.

The critics’ counter-reading of that standing must be given its strongest form, because the evenhandedness this series practises requires it and because the reading identifies a real tension. Opponents of the BJP argued through the 1990s, and historians of the movement have repeated the argument since, that Vajpayee’s moderation functioned as the acceptable face of a movement whose harder commitments remained unamended, a mask behind which the organisation pursued the programme the face disclaimed; the demolition of 6 December 1992, carried out by a mobilisation the party’s own campaign had fed, stood as their principal exhibit. The record of 1996 to 1998 does not refute the tension; it institutionalised it. The National Agenda for Governance was precisely the device by which the alliance partners insured themselves against the harder programme, in writing, for the life of the coalition, and the premium they charged, the shelving of the three signature commitments, was the price of the suspicion the critics articulated. Vajpayee’s achievement was to convert that suspicion from a veto into a price, and a price can be paid. After 1998 the conversion held for six years of coalition government. The doctrine needed the voice because the voice was the collateral; the institution the personality built outlasted the thirteen days that revealed its design.

A word is owed to the reader who suspects the argument of reading history backwards, of finding in a 1996 speech the blueprint of a 1998 outcome that had a hundred parents. The suspicion is healthy and the record answers it, because the correspondence between the speech and the settlement does not rest on the historian’s hindsight. It rests on the testimony of the interval: the party’s own councils named the isolation as the problem within weeks of the resignation; the negotiators of 1997 and 1998 priced partnership in the currencies the refusals of May 1996 had quoted; the common programme’s drafting rule, shared ground legislated and contested ground shelved, was the speech’s second principle reduced to a clause; and the partners who entered the alliance cited the conduct of the thirteen days as their security for entering it. Blueprints are not usually announced in advance by their architects on the floor of a parliament. This one was, which is why the case for the structural reading does not depend on believing that Vajpayee foresaw 1998 in detail. It depends on noticing that he stated, in May 1996, the conditions under which a party in his party’s position could govern India at all, and that the party which met those conditions was his own, twenty-two months later, with him at its head.

The Signal Reading, Adjudicated

The namable claim can be restated against the full record. The 1996 thirteen days mattered for what they signalled; the resignation speech of 28 May 1996 was the foundation document of the 1998 National Democratic Alliance. The first half is verified by the behaviour of every actor in the system: the regional parties repriced their partnership, the Congress repriced its outside backing, and the BJP rebuilt its strategy, its programme and its leadership pitch around the lessons of a fortnight. The second half is verified by correspondence: the three principles stated in the speech, legitimacy earned by conduct, government by negotiated programme, arithmetic subordinated to agreement, are the three load-bearing features of the alliance that took office on 19 March 1998. The footnote reading survives only by declining to look at 1998, and a reading of 1996 that cannot look at 1998 has surrendered the historian’s first duty, which is to follow consequences. The thirteen days were the shortest government of the Indian republic to their date. They were also the coalition era’s founding seminar, and the seminar’s lecture, delivered by a prime minister in the act of resigning, is still the cleanest statement of how power in that era was actually obtained.

Frequently Asked Questions

Q: What was the 1996 Vajpayee government?

The 1996 Vajpayee government was the first Bharatiya Janata Party government of India, sworn in on 16 May 1996 after the general election returned a hung Eleventh Lok Sabha in which the BJP, with 161 seats, was the single largest party and no formation commanded a majority. President Shankar Dayal Sharma invited Atal Bihari Vajpayee to form the government under the largest-party convention and required him to prove his majority on the floor of the House by 31 May 1996. The government governed for thirteen days, failed to attract the regional parties whose members it needed, and resigned on 28 May 1996 after the confidence debate, before the motion could be put to a vote it was certain to lose. It passed no legislation and is remembered for the doctrine its resignation speech announced.

Q: How long did Vajpayee’s first government last?

Vajpayee’s first government lasted thirteen days as a working ministry, from the oath of 16 May 1996 to the resignation he announced in the Lok Sabha and tendered to President Sharma on 28 May 1996, at the close of the two-day confidence debate. Because a resigning prime minister continues in a caretaker capacity until a successor is installed, reference listings of his tenure run from 16 May to 1 June 1996, the day H.D. Deve Gowda of the United Front took oath as prime minister. The thirteen days of the popular phrase count the interval between the oath and the resignation announcement, the period in which the government sought the majority it never found. It was, at its date, among the shortest tenures of any Indian prime minister who had faced the House.

Q: Why did Vajpayee resign in 1996?

Vajpayee resigned because the arithmetic of the Eleventh Lok Sabha could not be moved. The BJP’s 161 seats and its complete pre-poll alliance together commanded 187 members, 85 short of the 272 majority mark, and the parties holding the balance, the Telugu Desam Party, the Tamil Maanila Congress and the Asom Gana Parishad among them, refused to join or back a BJP government, citing the party’s Ayodhya record and the price their own electorates would exact. The remaining parties had organised the United Front in explicit opposition to the BJP’s inclusion and held, with Congress backing from outside, a demonstrated strength of some 332 members. Faced with certain defeat in the confidence vote, and unwilling to purchase members in defiance of the anti-defection law and of his party’s declared position, Vajpayee resigned from the floor of the House on 28 May 1996.

Q: What did Vajpayee say in his June 1996 speech?

The speech that closed the thirteen days was delivered in the Lok Sabha on 28 May 1996, in the government’s final hours before the caretaker interval that ended with Deve Gowda’s oath on 1 June, and it is catalogued in some references by that June transition. In his reply to the confidence debate, Vajpayee announced that the BJP would not purchase the majority it lacked, warned the combination preparing to succeed him that a government assembled by exclusion would find its birth difficult and its survival more difficult, and pledged that his party in opposition would back measures that served the country. He closed with the sentence quoted ever since: governments will come and go, parties will be formed and broken, but this country must endure and its democracy must remain immortal. The address is preserved in the Lok Sabha debates of 28 May 1996.

Q: How did the 1996 government lead to the 1998 NDA?

The thirteen days taught the BJP the three requirements of coalition power, and the party spent the twenty-two months after May 1996 meeting them. Its National Executive and National Council proceedings of 1996 and 1997 named the party’s isolation as the problem to solve; its state-level partnerships, with the Akali Dal in Punjab and the Shiv Sena in Maharashtra, supplied working models of shared government; and its negotiators converted the parties that had refused it in 1996, the Telugu Desam above all, into negotiators for 1998. The product was the National Agenda for Governance, a common programme that governed on shared ground and shelved the BJP’s three commitments its partners could not accept, Ayodhya legislation, Article 370 and a uniform civil code. The National Democratic Alliance that took office on 19 March 1998 was that preparation made visible.

Q: What was the post-Mandal political arithmetic?

The post-Mandal political arithmetic was the distribution of electoral power that followed the implementation of the Mandal Commission’s recommendations in August 1990, when the V.P. Singh government reserved 27 percent of central government posts for the Other Backward Classes and reorganised competition in the Hindi belt around caste constituencies. The Congress’s catch-all coalition fractured into parties of defined social blocs: backward-caste parties such as the Samajwadi Party and the Janata Dal and its successors, the Dalit constituency organised by the Bahujan Samaj Party, and the BJP’s upper-caste and middle-class base, alongside the established regional parties of the south and east. After 1989 no party could assemble a Lok Sabha majority from its own social bloc, and governments became aggregations of state-level verdicts negotiated between parties, which is the arithmetic that produced the hung house of 1996 and every coalition of the era.

Q: Did the BJP win the 1996 election?

The BJP won the 1996 election in the limited sense that it emerged as the single largest party, with 161 seats against the Congress’s 140, and thereby earned the President’s first invitation to form a government. It did not win a majority, it did not win the largest share of the vote, in which the Congress led with 28.80 percent against the BJP’s 20.29 percent, and it did not win the power to govern, since the remaining parties combined against it within days. The Election Commission’s statistical report records both the seat plurality that seated the government and the vote distribution that doomed it. The distinction between winning the election and winning the right to govern is the whole subject of the thirteen days: the BJP had won the first and had yet to learn the second in May 1996.

Q: Who invited Atal Bihari Vajpayee to form the government in 1996?

President Shankar Dayal Sharma invited Vajpayee to form the government, issuing the invitation on 15 May 1996 after the results of the general election had made the BJP the single largest party in a hung Lok Sabha. Sharma, a former Congress vice-president acting under the convention that had consolidated since 1989, gave the leader of the largest party the first opportunity to assemble a majority and attached a condition: the government would prove its strength on the floor of the House by 31 May 1996. No other formation had presented a demonstrated claim at the time of the invitation, the United Front and the Congress being still in negotiation, so the invitation followed size rather than prospects. Vajpayee was sworn in the next day, 16 May 1996, at Rashtrapati Bhavan, and the deadline structured every day of the government’s thirteen.

Q: What is a confidence motion, and why was the 1996 motion never put to a vote?

A confidence motion is the resolution by which a government asks the Lok Sabha to record that it commands the House’s confidence; a government that loses it must resign, and a government invited to office in a hung house is required by the President to seek one within a stated period. The Vajpayee government moved its motion, the debate ran across 27 and 28 May 1996 with some eleven hours of discussion recorded in a later PRS Legislative Research study of confidence motions, and every party whose members the government needed declared against it from the floor. With the count settled publicly before the division, Vajpayee used his reply to announce the government’s resignation, and a motion whose mover has resigned is not put to a vote. The 1996 motion therefore survives in the record as fully debated and never divided upon.

Q: Why did the Congress decline to head the government in 1996 despite its 140 seats?

The Congress declined because its leadership under P.V. Narasimha Rao read the same arithmetic as the BJP and drew the opposite institutional conclusion. A Congress-led government would have depended on the United Front parties assembled against the hawala-season Congress, would have carried five years of accumulated scandal into office, and would have faced destruction at its first floor test or slow strangulation thereafter. The party chose instead to back the United Front from outside, retaining the power to sustain or withdraw without the responsibility of office, and left the premiership to the Janata Dal’s H.D. Deve Gowda. The choice kept the BJP from power, which was its object, at a price the party paid over the following two years: withdrawal from Deve Gowda in April 1997, the destruction of the Gujral government that November, and defeat in the election of 1998.

Q: Which parties formed the BJP’s pre-poll alliance in the 1996 election?

The BJP contested the 1996 election in harness with a small set of state allies, principally the Shiv Sena in Maharashtra, the Samata Party of George Fernandes and Nitish Kumar in Bihar, the Shiromani Akali Dal in Punjab and the Haryana Vikas Party of Bansi Lal in Haryana, with seat adjustments in the states each partner commanded. The Election Commission’s returns credit the Bharatiya Janata Party’s alliance with 187 seats in aggregate against the BJP’s own 161, a combined strength that won the party the President’s invitation and left it 85 seats short of a majority. The alliance’s composition is itself evidence about the BJP’s position in 1996: its partners were parties of state or sectional strength, none of them able to bring a second region into the combination, and the great regional parties of the south, the Telugu Desam, the Dravida parties and the Tamil Maanila Congress, stood outside it, in the opposing organisation or in refusal.

Q: Which regional parties did the BJP court in May 1996, and why did they hold back?

The government’s soundings ran to the Telugu Desam Party of N. Chandrababu Naidu, holding 16 seats, the Tamil Maanila Congress of G.K. Moopanar, holding 20, and the Asom Gana Parishad of Prafulla Kumar Mahanta, holding 5, with lesser formations sounded behind them. Each held back for reasons rooted in its own electorate rather than in Delhi’s inducements. The Telugu Desam’s politics in Andhra Pradesh rested on a secular self-image and a rivalry with the Congress that BJP partnership could not serve; the Tamil Maanila Congress had just broken from the Congress on a platform of Tamil pride and could not define itself by adhesion to a party of the Hindi belt; the Asom Gana Parishad guarded an Assamese regionalism suspicious of every national formation. Behind each refusal stood the mobilisation around Ayodhya and the demolition of 6 December 1992, which had fixed the BJP in regional calculations as a partner whose embrace cost votes.

Q: What did the opposition argue in the confidence debate of 27 and 28 May 1996?

The opposition’s case rested on two arguments, each stated from the floor across the two days. The first was arithmetic legitimacy: a party holding 161 of 543 elected seats, opposed by the declared holders of nearly every remaining seat, could not claim the House’s confidence by the largest-party convention that had seated it, since the convention governed invitations and not majorities. The second was character: speakers from the Congress, the Janata Dal, the Left and the regional parties, Somnath Chatterjee of the CPI(M) prominent among them, held that the BJP’s record on Ayodhya and the demolition of December 1992 disqualified it from the trust of a plural polity, and that the moderation of its leader did not amend the commitments of the movement behind him. The treasury benches answered that a combination of thirteen parties and an outside backer, united only in exclusion, had no mandate either, a counter the succeeding two years kept in public view.

Q: What role did L.K. Advani play in the BJP’s 1996 strategy?

L.K. Advani was the BJP’s president through the thirteen days and the organisational half of the party’s dual leadership, Vajpayee supplying the parliamentary face and Advani the machine. Advani had announced, after being named in the Jain hawala chargesheets of early 1996, that he would not enter the Lok Sabha until the charge was cleared, so he directed the campaign and the government-formation effort from outside the House he had led for a decade. Within the party he carried the argument that the BJP had reached the ceiling of its own mobilisation and that power would come through coalition or not at all, a judgment whose authority derived from his standing as the architect of the mobilisation being declared insufficient, not least the rath yatra of 1990 and the Ayodhya campaign. After May 1996 he steered the party’s council deliberations toward the alliance strategy that his presidency, ending in 1998, handed to the coalition that took office that March.

Q: Did Vajpayee try to engineer defections to secure a majority in May 1996?

No attempt at engineering defections stained the thirteen days, and the absence was noticed across the divide. The anti-defection provisions of the Tenth Schedule, in force since 1985, had in any case outlawed the retail purchase of individual members, recognising only party-level splits and mergers, so the government’s search ran to parties as wholes and failed at that level. Vajpayee closed the question in his reply of 28 May 1996, announcing that the government would resign rather than assemble a majority by the methods the phrase horse-trading described, and the press coverage of the following day recorded opposition members conceding that no such attempt had been made. The refusal was both propriety and strategy: a purchased majority would have dissolved in weeks and destroyed the credential the government was banking, the demonstrated governability on which the courtship of 1996 to 1998 was built.

Q: What happened between Vajpayee’s resignation on 28 May 1996 and Deve Gowda’s oath on 1 June 1996?

Vajpayee tendered his resignation to President Sharma on the evening of 28 May 1996 and continued as caretaker prime minister while the alternative government completed its arrangements. The United Front’s constituents settled their internal questions in those four days: the premiership had been declined by V.P. Singh and made unavailable to Jyoti Basu by his own party’s central committee, and the Front’s steering committee converged on H.D. Deve Gowda, the Janata Dal chief minister of Karnataka, with Chandrababu Naidu as the Front’s convener managing the negotiation. The Congress formalised its decision to back the Front from outside. President Sharma then invited Deve Gowda, who resigned the Karnataka chief ministership, travelled to Delhi and took oath as prime minister on 1 June 1996, with eleven days allowed to prove his majority, which the Front’s numbers and the Congress’s backing comfortably supplied when the House met.

Q: What is political untouchability, and how did it describe the BJP’s position in 1996?

Political untouchability was the term Indian commentary of the 1990s used for the BJP’s exclusion from partnership: a party large enough to win seats across the northern and western states, and after 1996 to finish first among them, yet treated by every other formation as ineligible for coalition, office-sharing or outside backing. Christophe Jaffrelot’s The Hindu Nationalist Movement and Indian Politics of 1996 analysed the condition structurally, as the product of the party’s association with the Ayodhya mobilisation and the communal violence that attended it, which made association with the BJP a cost that regional parties computed against their own electorates. May 1996 was the condition’s purest demonstration: the BJP held 161 seats and the President’s invitation and could not obtain a single additional party. The adaptation of 1996 to 1998, which repriced partnership in interests rather than endorsement, is best understood as the party’s escape from that condition.

Q: Which parties that held back in 1996 joined or backed the BJP-led alliance in 1998?

The largest movement was the Telugu Desam Party of Chandrababu Naidu, convener of the United Front in 1996, which backed the Vajpayee government from outside in 1998 without taking ministerial office, receiving the Speakership of the Lok Sabha for its nominee G.M.C. Balayogi and a recognised voice in the alliance’s counsels. Around that movement the alliance of 1998 gathered parties that had stood outside the BJP’s reach in 1996: the AIADMK of J. Jayalalithaa in Tamil Nadu, the Biju Janata Dal founded by Naveen Patnaik in Orissa in December 1997, and the Trinamool Congress founded by Mamata Banerjee in January 1998, alongside the standing partners, the Shiv Sena, the Samata Party and the Akali Dal. The Tamil Maanila Congress and the Asom Gana Parishad, courted in 1996, held apart in 1998 as well, a selectivity that marks the courtship as a market in state interests rather than a general absolution.

Q: How did the thirteen days change Vajpayee’s personal standing in Indian politics?

The thirteen days converted Vajpayee from the BJP’s most acceptable leader into a national statesman whose credibility crossed the party divide, because the country watched him accept a doomed commission, govern without provocation for thirteen days and surrender office on schedule with a speech that opponents quoted with respect. His standing had older foundations, the foreign ministership of 1977 to 1979, the address to the United Nations General Assembly in Hindi, four decades in Parliament and a poet’s place in Hindi letters, but May 1996 gave those foundations a contemporary proof conducted under a deadline. The proof became a negotiable asset: the partners of 1998 accepted Vajpayee’s personal assurances as security for their risk in joining his alliance, and the alliance campaigned on his portrait rather than the party’s symbol across much of the country. The resignation, paradoxically, was the making of the prime ministerships that followed it.

Q: What did the 1996 results reveal about the BJP’s geographical limits?

The Election Commission’s state-level returns showed a party of the north and west confronting a map it could not complete. The BJP’s 161 seats were concentrated in Uttar Pradesh, Madhya Pradesh, Rajasthan, Gujarat, Delhi and, in harness with the Shiv Sena, Maharashtra, with footholds in Bihar, Orissa and Karnataka; across the south the party held almost nothing, in the east little more, and in the north-east nothing at all. A party so distributed could finish first in seats while remaining a regional power in a national costume, dependent for any majority on parties rooted in the territories where it could not campaign. The geographical limit was therefore also a strategic instruction, and the adaptation of 1996 to 1998 obeyed it: the alliance of 1998 was deliberately assembled as a federation of regional strengths, the AIADMK in Tamil Nadu, the Biju Janata Dal in Orissa, the Telugu Desam’s outside backing in Andhra Pradesh, each supplying the map the BJP lacked.