The post-Mandal party system is the most argued-over object in the study of Indian democracy, and the argument has produced three major readings that rarely meet on the same page. Yogendra Yadav reads the decades after 1990 as a deepening of democracy itself, a third wave in which citizens placed lowest in the caste order entered electoral politics as principals rather than as clients. Pradeep Chhibber reads the same decades as an institutional unravelling, the fragmentation of a once integrated party order into state-sized pieces that no national institution successfully reassembled for a quarter century. Christophe Jaffrelot reads them as the terrain on which Hindu nationalist mobilisation converted upper-caste anxiety and plebeian assertion into the Bharatiya Janata Party’s long ascent. This article defends one claim against the habit of choosing among the three: the three readings are complementary characterisations of a single system, because each one answers a different question about it. Yadav asks what changed for the governed, and answers with democratic content. Chhibber asks what changed for the architecture of competition, and answers with fragmentation. Jaffrelot asks what changed for the strategy of the most successful competitor, and answers with mobilisation. A reading that forces these into a single contest mistakes the map’s three layers for three rival maps.

The claim needs discipline before it needs evidence, because complementarity can become an excuse for never judging anything. The comparator reading defended here does judge. It holds that Yadav’s framework is the strongest of the three on the question it poses, that Chhibber’s framework is the strongest on its question, and that Jaffrelot’s framework is the strongest on its question, and it then specifies the points at which the frameworks genuinely collide rather than merely coexist. Three collisions matter. Yadav’s deepening thesis collides with Chhibber’s fragmentation thesis over whether the multiplication of parties after 1989 was a democratic achievement or an institutional loss. Chhibber’s fragmentation thesis collides with Jaffrelot’s mobilisation thesis over whether the Bharatiya Janata Party’s majority of 2014 ended the fragmented era or merely won inside it. Yadav’s deepening thesis collides with Jaffrelot’s mobilisation thesis over whether the post-2014 order, built partly on lower-caste votes, extends the third wave or turns against it. On each collision the electoral record can be made to testify, election by election and state by state, and this article puts that record first, before any framework is allowed to interpret it.

The evidence discipline throughout is the discipline the pillar has applied since its hub: settled facts stated plainly, disputed characterisations attributed to their holders at full strength. The settled facts here are unusually solid, because the Election Commission of India has published constituency-level statistical reports for every Lok Sabha and Vidhan Sabha election from 1989 onward, and the three frameworks all draw on the same reports, the same Lokniti survey programme at the Centre for the Study of Developing Societies, and the same post-election special issues of the Economic and Political Weekly for 2014, 2019 and 2024. The disputes begin only when scholars decide which rows of those tables carry the meaning of the period. That is precisely why the comparator reading is a finding and not a compromise: if three scholars reading identical tables produce three non-contradictory accounts, the explanation lies in the questions they brought to the tables.

The period’s two endpoints rhyme in a way the frameworks illuminate jointly. In 1989 a government was formed by a party of 143 seats resting on outside support from two blocs that despised each other, and it fell within a year. In 2024 a government was formed by a party of 240 seats resting on inside support from allies with state agendas of their own, and its programme was trimmed within weeks, on the budget’s regional allocations, on the lateral-entry advertisement withdrawn in August 2024, on the waqf legislation sent to a joint parliamentary committee. Between those endpoints, the system travelled from coalition improvisation to coalition convention to single-party majority and back to coalition dependence, and the travel is the strongest single argument against reading any one decade of it as its essence. Yadav’s waves explain why the electorate that produced both endpoints had been transformed between them. Chhibber’s architecture explains why governments at both endpoints were assembled from state-anchored parts. Jaffrelot’s mobilisation explains why the party at the centre of the 2024 assembly was the Hindu nationalist one rather than the Congress, the party at the centre of the improvisations of 1989. Remove any one dimension and the rhyme between the period’s first and last verdicts becomes inaudible.

One boundary must be drawn at the start. The Mandal moment itself, the commission, the 1990 implementation decision and the protests that followed it, belongs to companion studies in this pillar, including the three-reading adjudication framework that the hub article built for the whole caste and reservation sequence, and the three-phase mobilisation through which the implementation year unfolded. This article takes the consequences of that moment for party competition as its object. Its period runs from the general election of 1989, in which the Congress lost power for the second time in its history, to the general election of 2024, in which the Bharatiya Janata Party lost the single-party majority it had held for ten years and returned to coalition dependence. Between those two elections lies the entire career of the post-Mandal party system, and each framework is a theory of that career.

The Post-Mandal Party System and the Problem of Three Readings

The analytical problem is visible in a single contrast. Between 1952 and 1984, the Congress won a majority of Lok Sabha seats in eight of nine general elections, and a scholar describing the system could begin with one party and treat every other party as commentary. Between 1989 and 2009, no party won a Lok Sabha majority at all, and governments were built by the Janata Dal, then by the Bharatiya Janata Party at the head of the National Democratic Alliance, then by the Congress at the head of the United Progressive Alliance. A scholar describing that system had to begin with arithmetic. Yogendra Yadav, Pradeep Chhibber and Christophe Jaffrelot all began their accounts inside this contrast, and each of them located the meaning of the period in a different place. Yadav located it in the polling booth, in who was voting and winning. Chhibber located it in the party office, in how organisations linked citizens to the state and how those linkages thinned. Jaffrelot located it in the mobilisation camp, in the movement that learned to speak simultaneously in the idioms of caste assertion and Hindu unity.

Their biographies as scholars explain part of the divergence without reducing it to biography. Yadav built his framework as a psephologist attached to the Lokniti programme, which from 1996 onward ran National Election Studies that recorded the caste, class and locality of voters in samples large enough to track subaltern participation state by state. His raw material was the voter. Chhibber built his framework as a student of party organisations, first in Democracy without Associations, published by the University of Michigan Press in 1999, where his raw material was the party branch, the membership register and the absent intermediate associations of Indian public life. Jaffrelot built his framework as a historian of the Hindu nationalist movement, in The Hindu Nationalist Movement and Indian Politics, published in 1996 from his French doctoral research, where his raw material was the shakha, the pracharak network and the cadre diary of the Rashtriya Swayamsevak Sangh. Three kinds of archive produced three analytical dimensions, and the comparator reading begins by taking that seriously: a framework’s dimension is set by what its evidence can see.

The naming of the period already encodes the dispute. Calling the system post-Mandal privileges the August 1990 decision of the V. P. Singh government to implement the Mandal Commission’s recommendation of 27 percent reservation for Other Backward Classes in central government employment, a decision examined in the V.P. Singh government’s 1990 Mandal context. Yadav’s vocabulary accepts that privilege, because for him the implementation was the trigger of a participation revolution. Chhibber’s vocabulary resists it: his period break falls in 1989, with the end of Congress dominance, and Mandal appears in his work as one accelerant among several rather than as a founding event. Jaffrelot’s vocabulary splits the difference in an instructive way: his silent revolution is dated not to a cabinet decision but to the state elections of the 1990s, and his account of Hindutva treats 1990 as the year two mobilisations, Mandal and the Ram temple campaign, were launched against each other. The comparator reading keeps the post-Mandal label because the pillar’s architecture requires it, and flags the label itself as the first exhibit of the frameworks’ divergence.

What would count against the comparator reading? If the three frameworks made contradictory predictions about the same observable quantity, they could not all be right, and choosing among them would be unavoidable. Section by section, this article tests for exactly that. On turnout by caste, Yadav’s framework predicts a narrowing and reversal of the participation gap between upper castes and Other Backward Classes in the northern states during the 1990s; neither rival framework predicts the opposite. On the number of parties winning seats and the duration of governments, Chhibber’s framework predicts sustained fragmentation from 1989 to 2014; neither rival predicts consolidation in that window. On the Bharatiya Janata Party’s vote share and seat trajectory, Jaffrelot’s framework predicts a long, regionally uneven ascent punctuated by the counter-mobilisation of caste parties; Yadav’s framework qualifies the interpretation of that ascent without denying its numbers. Where real contradictions appear, in the pairwise tensions, this article says so and adjudicates them on the record rather than dissolving them into complementarity. That is the standard the hub set for this pillar through the three-reading adjudication framework, and the standard this article accepts for itself.

The Electoral Record from 1989 to 2024: What Any Reading Must Explain

Every framework must first survive contact with the seat record, so the record deserves a section of its own, stated without adjectives. In the general election of November 1989, the Congress won 197 seats on 39.5 percent of the vote, the Janata Dal won 143, the Bharatiya Janata Party won 85, and the two communist parties won 45 between them; V. P. Singh formed a government supported from outside by both the Bharatiya Janata Party and the Left, an arrangement whose fragility recalled the post-1977 Janata coalition’s institutional fragility. In 1991, in an election interrupted by the assassination of Rajiv Gandhi on 21 May, the Congress won 244 seats, the Bharatiya Janata Party 120 and the Janata Dal 59. In 1996 the Bharatiya Janata Party became the largest party with 161 seats, Atal Bihari Vajpayee’s first government fell after thirteen days when the regional parties he courted declined to sustain it, and two United Front governments under H. D. Deve Gowda and I. K. Gujral governed with Congress support from June 1996 to March 1998. In 1998 the Bharatiya Janata Party won 182 seats and built the National Democratic Alliance; in 1999 it won 182 again, the alliance won a working majority, and Vajpayee governed a full term. In 2004 the Congress won 145 seats against the Bharatiya Janata Party’s 138 and governed for a decade through the United Progressive Alliance; in 2009 the Congress rose to 206 while the Bharatiya Janata Party fell to 116.

The second half of the record reverses the first half’s direction. In 2014 the Bharatiya Janata Party won 282 seats on 31.3 percent of the vote, the first single-party majority since 1984, and the Congress fell to 44 seats on 19.5 percent, its lowest tally in either measure; the analysis of that break with the post-Mandal arithmetic is pursued in the post-Mandal arithmetic break. In 2019 the Bharatiya Janata Party rose to 303 seats on 37.4 percent of the vote, a consolidation examined in the 2019 verdict’s structural confirmation, and the Congress recovered only to 52. In 2024 the Bharatiya Janata Party fell to 240 seats on 36.6 percent, below the majority mark of 272, and Narendra Modi’s third government rested on the Telugu Desam Party’s 16 seats and the Janata Dal (United)’s 12, among smaller partners. The Congress rose to 99 seats, the Samajwadi Party won 37 in Uttar Pradesh alone, and the opposition INDIA coalition took the combined opposition tally past 230. Governments between 1989 and 2014 lasted, in sequence, roughly eleven months, ten months, five years in two instalments, two years in two instalments, thirteen months, one full term under the Alliance, and ten years under the United Progressive Alliance. Durability arrived before majority did, a fact the fragmentation framework has to explain rather than merely cite.

Turnout and participation complete the record’s skeleton. National turnout stood at 59.7 percent in 1989, dipped to 56.7 percent in 1991, recovered to 62.0 percent in 1998, fell to 58.0 percent in 2009, and then rose to 66.4 percent in 2014 and 67.4 percent in 2019 before easing to 65.8 percent in 2024. The composition of that turnout changed more than its volume, and the compositional evidence is the terrain of Yadav’s framework, so the details belong to his section. One compositional fact belongs here because all three frameworks must digest it: the social profile of the Lok Sabha itself altered. Jaffrelot’s tabulations from the Election Commission’s candidate records show that among members elected from the Hindi belt states, the upper-caste share fell from roughly one half in the 1950s and 1960s to under one third by the late 1990s, while the Other Backward Classes share rose from roughly one in nine members in 1984 to roughly one in four by 1996. No framework denies these numbers. They disagree, entirely, about what the numbers did to Indian democracy.

How long did the two United Front governments of 1996 to 1998 last?

The United Front experiment ran for just under twenty two months across two premierships. H. D. Deve Gowda took office on 1 June 1996 and resigned on 21 April 1997, after the Congress withdrew support. I. K. Gujral succeeded him on 21 April 1997 and resigned on 28 November 1997, when the Congress withdrew support again.

Gujral then continued as caretaker prime minister until 19 March 1998, when the Vajpayee government was sworn in after the mid-term election, closing the period in which outside support functioned as a governing instrument.

The state record matters as much as the parliamentary one, because the post-Mandal system was assembled in the states before it was named at the centre. Uttar Pradesh held five assembly elections between 1989 and 2002 without returning any party to power, and its politics ran through the Samajwadi Party, founded by Mulayam Singh Yadav in October 1992, and the Bahujan Samaj Party, founded by Kanshi Ram in April 1984, whose leader Mayawati became chief minister in June 1995 with Bharatiya Janata Party support. Bihar ran through the Janata Dal of Lalu Prasad Yadav, chief minister from March 1990, and after the Janata Dal’s successive splits through his Rashtriya Janata Dal, founded in July 1997, and the Janata Dal (United) of Nitish Kumar. Tamil Nadu alternated between the Dravida Munnetra Kazhagam and the All India Anna Dravida Munnetra Kazhagam without interruption from national parties. West Bengal’s Left Front, in office since 1977, won its seventh consecutive assembly election in 2006 and then lost to Mamata Banerjee’s Trinamool Congress in 2011. Any reading of the national system that cannot see these state systems is reading the shadow rather than the object, a point on which, unusually, all three frameworks agree in practice even where their theories differ.

The upper house tells the same story on a lag. Rajya Sabha composition, renewed by thirds from the state assemblies, transmitted the post-2014 state-level balance to Parliament years after the Lok Sabha had registered the national verdicts, which is why the majority governments of 2014 to 2024 legislated through a chamber their alliance did not control for most of the period, and why ordinances, joint sittings and the classification of bills as money bills became recurring instruments of the era’s law-making. The institutional friction is Chhibber’s dimension expressing itself inside the constitution’s own machinery: a fragmented federal base will surface somewhere in the legislative process, in coalition bargaining if the lower house is fragmented, in upper-house arithmetic if it is not.

Yogendra Yadav and the Three Waves of Democratic Deepening

Yogendra Yadav’s framework, assembled across essays from the mid 1990s and gathered in Making Sense of Indian Democracy, published by Permanent Black in 2020, periodises Indian democracy as three waves of deepening. The first wave, from 1947 into the 1960s, built the formal architecture: universal adult franchise adopted by the Constituent Assembly in 1949 against the advice of many contemporaries, a working Election Commission under Sukumar Sen conducting the first general election of 1951 to 1952, and courts, legislatures and a free press operating from the republic’s first day. The second wave, running from the late 1960s through the 1980s, widened the social base of the formal architecture from inside the Congress organisation itself, as the party’s state-level leaderships passed to Other Backward Classes politicians in the southern and western states and as opposition victories in 1967 and 1977 proved that incumbents could be removed by ballot. The third wave, from 1989 and 1990 onward, carried that widening into the northern heartland and into the lower reaches of the caste order: Other Backward Classes, Scheduled Castes and Scheduled Tribes began to vote in proportions matching or exceeding upper-caste participation, to win seats in proportions approaching their population shares, and to set the agenda of state politics through parties of their own.

The framework’s engine is a distinction between formal democracy and democratic deepening. Formal democracy is a property of institutions and can be dated to a founding; deepening is a property of participation and representation, and it arrives in instalments carried by social groups entering politics in sequence. Yadav developed the sequence’s later instalment in his essay “Understanding the Second Democratic Upsurge”, published in 2000 in the volume Transforming India, edited by Francine Frankel, Zoya Hasan, Rajeev Bhargava and Balveer Arora. There he named the post-1990 surge of lower-caste participation in the northern states the second democratic upsurge, distinguishing it from a first upsurge in the 1960s and 1970s, when Other Backward Classes had entered electoral politics in force in Tamil Nadu, Karnataka, Maharashtra and Gujarat. The northern lag, and its end, is the framework’s central empirical claim: the region containing roughly two fifths of the Lok Sabha had kept its lower-caste citizens as vote banks for upper-caste-led parties a generation longer than the south and west, and Mandal ended that arrangement.

Three kinds of evidence carry the claim. The first is turnout by social group, drawn from the Lokniti National Election Studies. In the 1996 general election, survey estimates placed Other Backward Classes turnout in Uttar Pradesh and Bihar at levels equal to or above upper-caste turnout for the first time in the survey record, reversing a gap that the studies of the 1970s had measured at ten points or more. The second is descriptive representation, the caste profile of legislators, where Jaffrelot’s own tables supply Yadav’s numbers: a transfer of seats from upper castes to Other Backward Classes across the 1990s in every Hindi belt state, steepest in Bihar and Uttar Pradesh. The third is the partisan vehicle: the upsurge travelled through parties led by the groups it mobilised, the Samajwadi Party and the Bahujan Samaj Party in Uttar Pradesh, the Janata Dal and its successors in Bihar, the Dravida parties in Tamil Nadu a generation earlier, so that the deepening altered not only who voted but who could plausibly govern. Yadav’s 2009 essay on the third wave’s maturing, and his post-2014 essays, extended the argument to argue that plebeian assertion had become a permanent feature of the system rather than a Mandal-era episode, surviving even the Bharatiya Janata Party’s majorities, because the party was itself compelled to recruit, promote and reward lower-caste leaderships on a scale its 1980s predecessor never attempted.

The framework’s scope is deliberate and its author has been candid about it. Yadav’s question is what democracy means for the citizen placed at the bottom of a hierarchical order, and his measure of the system is the depth of that citizen’s effective membership in it. He does not offer a theory of cabinet stability, of party finance or of legislative productivity, and his framework would not be weakened by evidence that post-1990 governments were unstable, because instability is not his variable. This candour is what makes the comparator reading possible. It also exposes the framework’s vulnerable flank: a deepening measured by participation and descriptive representation can coexist with stagnant policy outcomes for the groups whose participation deepened, and critics from several directions, including scholars of the Bahujan movement itself, have pressed the point that presence in the legislature is not the same as power over the budget. Yadav’s answer, given in his 2020 collection, is that the sequence matters more than the snapshot: political equality arrived first in India, before social and economic equality, and the political arrival is the instrument with which the later equalities get demanded. The comparator reading records that answer as a claim about sequence, testable against decades rather than against a single parliament.

Two neighbouring accounts test the framework’s edges and are worth distinguishing from it, because careless summaries merge all three. Kanchan Chandra’s Why Ethnic Parties Succeed, published by Cambridge University Press in 2004, explains the same lower-caste party success by a colder mechanism: in a patronage democracy, where the state is the principal source of livelihood and services are delivered discretionarily, voters back ethnic parties when the ethnic headcount makes victory plausible and when co-ethnic office-holders can be monitored for delivery. Chandra’s voter is calculating rather than asserting, and her account predicts success for precisely the parties Yadav celebrates, for reasons his framework does not require. The two accounts divide the evidence in a familiar comparator pattern: Chandra’s explains the choice among caste parties on offer in a given constituency, Yadav’s explains why such parties existed to be chosen. A second test comes from within the Bahujan tradition itself: Gopal Guru’s writings on Dalit politics, including his essays on the limits of descriptive representation, ask whether the entry of the oppressed into legislatures altered the terms on which they were governed, or merely the personnel of their governing. Yadav’s sequence argument is the framework’s reply, and the dispute between them, unlike the framework disputes adjudicated below, is a dispute about the destination of the same road, which both scholars agree the country travelled.

The Second Democratic Upsurge in the States

Frameworks earn their keep in the states, and Yadav’s upsurge thesis was built state by state. Uttar Pradesh supplies its purest case. In the assembly elections of 1989 and 1991, the Janata Dal and then the Bharatiya Janata Party competed for a lower-caste electorate still addressed through intermediaries; by the assembly election of 1993, fought in the shadow of the Babri Masjid demolition of 6 December 1992, the Samajwadi Party and the Bahujan Samaj Party together won 176 seats and formed a government under Mulayam Singh Yadav with Mayawati’s party in support, the first Uttar Pradesh government in which both coalition partners were led by politicians from the mobilised castes themselves. The Bahujan Samaj Party’s subsequent trajectory, from 67 seats in 1996 to a full majority of 206 seats in 2007 under Mayawati, is the single strongest state-level exhibit for the deepening thesis: a party founded in 1984 to represent Scheduled Castes, led by a woman from the Chamar caste, governed India’s largest state alone. That its 2007 victory rested on a deliberate broadening to Brahmin candidates and voters, which Mayawati herself named a strategy of sarvajan, or all-community, appeal, complicates but does not refute the thesis; the broadening was conducted from a position of Dalit-led strength that the pre-1990 system never offered.

Bihar supplies the second case, and the harder one. Lalu Prasad Yadav took office in March 1990 as the Janata Dal’s chief minister, and his fifteen years at the head of Bihar’s governments, his own until 1997 and his wife Rabri Devi’s thereafter, delivered the upsurge’s representational content with unmistakable clarity: Yadav and other backward-caste legislators displaced the upper-caste leaderships that had run Bihar’s Congress and its governments since 1947, and the state’s administrative and police leaderships were pressed, often crudely, to reflect the shift. The same fifteen years delivered the critics’ evidence in equal measure: Bihar’s road, power and school indicators fell further behind the national averages, the fodder scam prosecutions that began in 1996 documented the diversion of roughly 950 crore rupees from animal husbandry treasuries, and the state’s per capita income gap with the rest of India widened across the decade. Yadav’s framework absorbs the Bihar case as the price and proof of a sequence in which dignity and presence precede administrative performance. Its rivals read the same case as a warning that representation without institutional capacity decays into patronage. The comparator reading does not split this difference; it assigns the two readings to their dimensions. As a claim about who held power in Bihar in 1995 compared with 1985, the upsurge thesis is simply true. As a claim about what that power delivered, the thesis is incomplete, and it was left to Bihar’s voters to deliver the institutional verdict, which they did in November 2005 by electing Nitish Kumar’s alliance on an explicit platform of roads, electricity and law enforcement.

The southern states discipline the framework’s chronology. Karnataka’s Other Backward Classes politics ran through Devaraj Urs, chief minister from 1972 to 1977, whose land reforms and backward-classes commissions made the state an early site of the first upsurge; Tamil Nadu’s Dravidian parties had displaced the Congress by 1967 and made backward-caste leadership of the state routine a full generation before Mandal. This geography is why Yadav speaks of waves rather than of a single revolution: the deepening moved region by region, caste bloc by caste bloc, through the federal system’s separate electoral arenas, and a national aggregate conceals the movement’s sequence. The comparator reading underlines the consequence for Chhibber’s and Jaffrelot’s frameworks: any national story about the post-1990 system is a weighted sum of state stories that began in different decades, and the frameworks differ partly because each scholar’s exemplary states differ, Karnataka and Tamil Nadu for the early wave, Uttar Pradesh and Bihar for the second, Gujarat and Madhya Pradesh for the Hindu nationalist consolidation.

Did women’s turnout reach parity with men’s turnout in the 2019 general election?

Yes, on the Election Commission’s published figures, and the convergence took six decades. In 1962, turnout among women stood near 46.6 percent against roughly 63.3 percent among men, a gap of about seventeen points. By 2014 the gap had narrowed to about one and a half points, and in 2019 women voted at the higher rate.

The precise 2019 figures, 67.18 percent for women against 67.01 percent for men, closed a narrowing that the Election Commission’s statistical reports had tracked through every decade: a gap of about four points remained in 2009, after standing near ten points in the elections of the 1990s.

The upsurge thesis also carries a prediction about party behaviour that the later record tests. If lower-caste participation and representation became structurally central after 1990, then every ambitious party, including a party of Hindu nationalist consolidation, would have to compete for lower-caste votes with lower-caste candidates and leaders, rather than merely enduring lower-caste parties as opponents. The Bharatiya Janata Party’s practice after 2013, examined in Jaffrelot’s section below, meets that prediction almost point for point: the projection of Narendra Modi’s own Other Backward Classes origin in the 2014 campaign, the cultivation of non-Yadav Other Backward Classes and non-Jatav Scheduled Castes as distinct electoral constituencies, the appointment of Keshav Prasad Maurya to lead the Uttar Pradesh unit in 2016, and the elevation of Droupadi Murmu, a Santhal Adivasi, to the presidency in 2022. Yadav himself noted the pattern in his post-2019 essays, reading it as the third wave’s logic working through its apparent victor. Whether that reading survives scrutiny belongs to the pairwise adjudication; here it is enough to register that the upsurge framework, alone among the three, treats the composition of the winning side as primary evidence.

Pradeep Chhibber and Fragmentation after the Congress System

Pradeep Chhibber’s framework begins where Rajni Kothari’s ended. Kothari’s account of the Congress System, set out in his 1964 essay and elaborated in Politics in India in 1970, described a one-party-dominant order in which the Congress contained the system’s principal factions inside itself and opposition parties functioned as parties of pressure around a consensus they could not displace; the anatomy of that order is traced in the Congress System’s four-component framework. Chhibber’s Democracy without Associations, published in 1999, asked what replaced that order once Congress dominance collapsed, and answered with a deliberately unglamorous word: fragmentation. The party system that emerged after 1989 was fragmented in three senses at once. The vote fragmented, as the combined share of the two largest national parties fell and state parties multiplied. The legislature fragmented, as the number of parties winning Lok Sabha seats rose from roughly a score in the 1980s to more than thirty in 1996 and near forty by 2009. Authority fragmented, as state governments led by regional parties became the effective sovereigns of their party systems, leaving the national arena an aggregate of state contests rather than a contest of its own.

The title of the 1999 book carries the framework’s causal claim. Indian parties, Chhibber argued, developed without the dense web of voluntary associations, unions, professional bodies and civic organisations that linked parties to society in the classic European cases, and which in those cases aggregated interests, recruited members and disciplined programmes. In their absence, Indian parties linked themselves to voters through the state itself: through patronage, through the discretionary distribution of public employment, licences, contracts and welfare, and through direct appeals to ascriptive blocs whose leaders could deliver enumerated support. A party system built on such linkages fragments predictably, because the linkages are divisible: any leader with a caste bloc, a district network or a ministry’s patronage can found a party, and the split costs little where no associational membership holds the founder inside the parent organisation. The Janata Dal’s disintegration into the Samajwadi Party in 1992, the Rashtriya Janata Dal in 1997, the Biju Janata Dal in 1997, the Janata Dal (Secular) in 1999 and the Janata Dal (United) is the framework’s textbook sequence, four successor parties and a rump from one parent within a decade, each split following a leader’s state base rather than a programme.

Chhibber’s second major statement, Ideology and Identity, written with Rahul Verma and published by Oxford University Press in 2018, complicated the first in a way the comparator reading must register honestly. Drawing on Lokniti survey evidence, Chhibber and Verma argued that Indian party competition is structured by two ideological dimensions after all: a politics of statism, over how far the state should intervene in the economy and regulate social life, and a politics of recognition, over whether the state should extend special protections to minorities and lower castes. Parties occupy distinguishable positions on both dimensions, voters sort among them partly on those positions, and the apparent chaos of fragmentation conceals an ideological order. The 2018 book does not retract the 1999 argument; it relocates it. Fragmentation remains the system’s institutional form, but the form is inhabited by ideological competition of a recognisable kind. Critics who cite the later Chhibber against the earlier one are therefore citing a refinement the author built himself, and the comparator reading treats the two books as one framework in two editions: the architecture fragmented, the competition within the architecture ideological.

The framework’s distinctive evidence is institutional and measurable. Its preferred statistics are the effective number of parties, a measure that weights parties by their seat or vote shares, which rose from under three in the elections of the early 1980s to above five in the Lok Sabha elections of the 1990s; the share of seats held by parties confined to a single state, which rose across the same decades; the tenure and turnover of governments; and the frequency of party splits and mergers recorded by the Election Commission’s recognition lists. Its preferred cases are organisational: the decay of the Congress’s internal elections after the 1970s, the cadre discipline of the communist parties and the Bharatiya Janata Party read as exceptions that prove the associational rule, and the personalist succession patterns of the regional parties, where leadership passes by family designation in the Dravida Munnetra Kazhagam, the Samajwadi Party, the Rashtriya Janata Dal, the Shiromani Akali Dal and the National Conference alike. Where Yadav sees democratisation in the rise of such parties, Chhibber sees the substitution of lineage for organisation, and the pairwise adjudication below takes that collision directly.

What the framework deliberately brackets is content. Chhibber’s question is how competition is organised, not what it delivers or whom it emancipates, and his framework would not be damaged by evidence that fragmented parliaments passed major legislation, though he would reply that the Anti-Defection Act of 1985 and the authority of party whips, not the party system’s health, explain legislative throughput. Nor does the framework require that fragmentation be permanent: a realignment that rebuilt national linkages would register in his measures as a falling effective number of parties and a rising two-party vote share, and the record after 2014, on which the Bharatiya Janata Party and the Congress together took 57 percent of the vote in 2019, is data for his framework and not an embarrassment to it. The embarrassment, if any, is subtler: whether a system whose largest party wins 240 seats and governs by coalition, as in 2024, counts as fragmented or as consolidated turns out to depend on thresholds the framework leaves implicit, and the adjudication section presses exactly there.

Coalitions as Architecture: 1989 to 2014 and the Majority Test

The institutional history of the post-Mandal decades reads, in Chhibber’s dimension, as a sequence of attempts to govern a fragmented legislature. The National Front government of 1989 rested on outside support from two mutually hostile blocs and fell in November 1990 when the Bharatiya Janata Party withdrew support during the Rath Yatra crisis, after eleven months. The Chandra Shekhar government that followed rested on Congress sufferance and fell in four months. The Congress government of P. V. Narasimha Rao, formed in 1991 with 244 seats, survived a full term by converting minorities into working majorities through by-election gains and floor management, and carried the 1991 economic liberalisation through the same fragmented house that sceptics had declared ungovernable. The United Front governments of 1996 to 1998, chronicled in the record section, fell twice to Congress withdrawals. The National Democratic Alliance then stabilised the form: the alliance of 1998 fell in thirteen months by a single vote in April 1999, but the enlarged alliance of 1999 governed five years, and the United Progressive Alliance governed ten, with its second term surviving the Left’s withdrawal of support in July 2008 over the civil nuclear agreement by winning a trust vote with Samajwadi Party assistance.

Two lessons follow for the framework. First, the coalition era’s instability was front-loaded: between 1989 and 1999 India held five general elections, and between 1999 and 2024 it held five more in twenty five years, a rhythm indistinguishable from the pre-1989 norm. Fragmentation of the seat distribution persisted throughout, while governability was recovered by alliance craft, by the anti-defection regime’s raising of the cost of exit, and by the convention, visible from 1999, that pre-electoral alliances with common programmes outlast post-electoral improvisations. A fragmentation framework that predicted chronic instability from seat shares alone would have been falsified by 2004; the stronger version of Chhibber’s argument, which treats fragmentation as a condition to be managed rather than a sentence of instability, is the version the record supports. Second, the state’s party systems, on which the national aggregate rests, stabilised earlier than the national one: most large states settled by the late 1990s into two-front or two-party contests, Tamil Nadu between the Dravida fronts, Kerala between the Left Democratic Front and the United Democratic Front, Uttar Pradesh among three and then effectively two competitors after 2017. The national system looked fragmented partly because it summed unlike state duopolies, a point developed by Chhibber himself in his collaborative work on party aggregation and by E. Sridharan in his studies of coalition politics in the same decades.

Chhibber’s framework also identifies its own exceptions, and the exceptions have grown more instructive with time. The cadre parties, the communists in their strongholds and the Bharatiya Janata Party through its Sangh linkage, possessed the associational depth his Indian norm lacks: membership registers, full-time organisers, affiliated unions, student and women’s wings, and internal promotion ladders that a founder could not bequeath. In the 1999 argument these parties were the deviation; by the 2010s the deviation had won the system, and a fair statement of the framework must record that the associational exception out-organised the associational norm across the majority decade. The Congress’s presidential election of October 2022, in which Mallikarjun Kharge defeated Shashi Tharoor in a contested vote of the party’s electoral college, the first such contest in over two decades, enters the same ledger from the other side: an attempt, late and partial, to rebuild an associational fact inside a party whose fragmentation-era habits the seventh pillar of this series has documented. Chhibber’s measures can score both events, a falling effective number of parties at the top, an unusual internal election inside a declining incumbent, without revising the framework’s premise, because the premise concerns the default condition of Indian parties, and defaults are confirmed, not refuted, by the effort required to escape them.

Did the 2014 majority end the coalition era at the centre?

In form it suspended the coalition era, and in substance it qualified it. The Bharatiya Janata Party’s 282 seats in 2014 delivered the first single-party Lok Sabha majority since 1984, and its 303 seats in 2019 deepened that position, so no coalition arithmetic decided the government’s survival in that decade.

The government nonetheless retained its National Democratic Alliance partners in cabinet, governed several states through alliances, and depended on friendly parties in the Rajya Sabha, where it lacked a majority for most of the period; and in 2024 the party’s fall to 240 seats restored coalition dependence in the Lok Sabha itself.

The majority decade therefore functions in this article as a controlled test of the fragmentation framework. If fragmentation described only the absence of single-party majorities, 2014 would have ended the post-Mandal system and the framework would be a period piece about 1989 to 2014. If fragmentation describes the underlying organisation of competition, state-anchored parties, divisible linkages, personalist successions, personalist campaign organisations, then the majority decade shows a national party winning inside the fragmented architecture by mastering its state units one arena at a time, and 2024 shows the architecture reasserting itself the moment the mastery slipped in Uttar Pradesh, Maharashtra and West Bengal. Chhibber’s own later writing leans to the second description, and the comparator reading adopts it as the framework’s defensible form: fragmentation as the system’s geology, majorities as weather. The cost of the metaphor is that geology changes slowly and explains timing badly, which is why the framework needs Jaffrelot’s, a theory of the actor who learned to operate the geology, to complete the comparator set.

Christophe Jaffrelot and the Silent Revolution

Christophe Jaffrelot’s framework rests on a book whose title became the period’s shorthand. India’s Silent Revolution: The Rise of the Lower Castes in North Indian Politics, published by Hurst and Columbia University Press in 2003, documented the transfer of political power in the northern states from upper castes to Other Backward Classes and Scheduled Castes across the twentieth century, at a pace set not by insurrection but by elections: the slow conversion of numerical strength into candidacies, ministries and chief ministerships. The revolution was silent, in his usage, because it proceeded through institutional channels and attracted little commentary while it happened, and it was a revolution because the social identity of the ruling stratum changed. Jaffrelot dated its northern completion to the 1990s, when the Mandal implementation, the rise of the Samajwadi Party and the Bahujan Samaj Party, and the Janata Dal’s Bihar governments together ended the upper castes’ hold on the Hindi belt’s legislatures, which his tables trace assembly by assembly and parliament by parliament from 1952 onward.

The same scholar wrote the period’s other master narrative, and the framework’s distinctiveness lies in holding the two together. In The Hindu Nationalist Movement and Indian Politics, published in 1996, and across the essays that led to Modi’s India, published by Princeton University Press in 2021, Jaffrelot traced the Hindu nationalist movement’s response to plebeian assertion. The response, in his account, had two registers. The first was counter-mobilisation: the Ram temple campaign, Lal Krishna Advani’s Rath Yatra of September and October 1990, and the kar seva mobilisations of 1990 and 1992 appealed to a Hindu unity that would override caste division precisely when caste division had acquired parties and ministries; the temple movement’s social base in its first phase drew disproportionately on upper castes anxious at the Mandal settlement, a relationship traced in the three-phase mobilisation of the implementation year. The second register was accommodation: from the mid 1990s, and decisively after 2013, the Bharatiya Janata Party pursued what its strategists called social engineering, the systematic recruitment of Other Backward Classes and Scheduled Caste voters through candidates, alliances and welfare delivery, under a leadership that foregrounded Narendra Modi’s own backward-caste origin. Jaffrelot’s framework thus reads the post-Mandal system as a dialectic: caste assertion created both the system’s democratising content and the constituency for a majoritarian strategy that promised to transcend caste in the name of the Hindu whole.

The framework’s evidence is prosopographical and ethnographic as much as electoral. Jaffrelot counts legislators by caste across seven decades, a labour no other scholar of the period matched, and he reads the internal literature and cadre structure of the Rashtriya Swayamsevak Sangh, founded in 1925, whose pracharak system supplied the Bharatiya Janata Party with organisers in districts where the party had no electoral history. His framework consequently sees things the other two miss: the transfer of the Sangh’s second-rung leadership to Other Backward Classes recruits during the 1990s, the party’s patient construction of caste-specific associations for communities such as the Nishads, Lodhs and Nonia in the northern states, and the long apprenticeship of the 2014 victory in state campaigns from Madhya Pradesh in 2003 to Bihar in 2005. Where Yadav counts voters and Chhibber counts parties, Jaffrelot counts people in offices, and the three counts are the comparator reading’s three dimensions made concrete.

His characterisation of the post-2014 order is the framework’s most contested element and must be rendered at full strength. In Modi’s India, Jaffrelot argues that the conjunction of Hindu nationalist ideology, a plebiscitary style of leadership that addresses voters over the heads of intermediaries, and the capture or weakening of countervailing institutions produced a system he names ethnic democracy: formal democratic procedures continue, elections remain competitive and are regularly lost by the ruling party in the states, but the majority community’s primacy is asserted as the regime’s principle and minorities participate on sufferance. Critics dispute the label from two sides: scholars who stress the ruling party’s repeated state-level defeats and its 2024 dependence on allies argue that competitive federalism contradicts the closure the label implies; scholars sympathetic to the government dispute the characterisation of its principle. The comparator reading does not adopt the label. It records the label’s location in the framework: it is Jaffrelot’s answer to his own question, what the mobilisation strategy did to the regime’s character once it won, and it stands or falls on institutional evidence of a kind his section of this article can present but not exhaust.

The framework’s blind spot is symmetrical to Yadav’s. A reading centred on the Hindu nationalist movement’s strategy can understate how much of the post-Mandal transformation proceeded independently of that movement: the Dravidian system’s consolidation owed nothing to it, the Left Front’s long rule and fall in West Bengal turned on agrarian and industrial questions internal to Bengal, and the Congress’s organisational decay, chronicled in this series’ seventh pillar, began a generation before Mandal. Jaffrelot’s own state chapters concede much of this, which is why the comparator reading treats his framework as a dimension and not as a total theory: it explains the period’s most consequential party and the strategic field its rise created, and it is silent, honourably, about dimensions it does not claim.

Hindu Nationalist Mobilisation as Strategy: The Long Ascent in the States

The seat trajectory of the Bharatiya Janata Party, stated plainly, is the ascent the framework must explain: 2 seats in 1984, 85 in 1989, 120 in 1991, 161 in 1996, 182 in 1998 and 1999, 138 in 2004, 116 in 2009, 282 in 2014, 303 in 2019, 240 in 2024. No other party of the post-Mandal period describes such a curve, and the curve’s inflections map onto the strategy’s phases. The 1989 and 1991 rises rode the temple mobilisation and the Ayodhya campaign’s polarisation of the northern electorate. The 1996 plurality converted movement strength into largest-party status without the allies that government required, and the thirteen-day government’s fall taught the party the coalition discipline it practised from 1998, moderating its programme in the National Democratic Alliance’s common agenda and accepting regional leaderships it could not control. The 2004 and 2009 defeats then mark the phase in which the temple constituency had been spent and the accommodation constituency not yet built: the party’s vote share stagnated in the low twenties, its state governments in Madhya Pradesh, Chhattisgarh and Gujarat held, but its national ceiling held lower.

The state record explains the ceiling and its breaking better than any national slogan. In Gujarat, the party governed from 1995 and, under Modi’s chief ministership from October 2001, fused development administration with Hindu nationalist organisation into the template later nationalised. In Madhya Pradesh, Shivraj Singh Chouhan’s governments from 2005 built a welfare-paternalist variant that held the state for eighteen years against the Congress. In Bihar, the party grew as Nitish Kumar’s junior partner from 2005, absorbing Other Backward Classes and Extremely Backward Classes voters through the alliance’s caste arithmetic rather than against it. In Uttar Pradesh, the party’s revival came last and fastest: reduced to 10 Lok Sabha seats in 2009, it won 71 of the state’s 80 seats in 2014 and 312 of 403 assembly seats in 2017, on a coalition of upper castes, non-Yadav Other Backward Classes and non-Jatav Scheduled Castes assembled constituency by constituency, with the Samajwadi Party and the Bahujan Samaj Party retaining their core blocs but losing the peripheries that decide first-past-the-post contests. The pattern repeated with local variations in Haryana in 2014, in Assam in 2016, in the North East through alliance brokerage thereafter, and partially in West Bengal in 2019 and 2021, where the party rose from 2 assembly seats in 2016 to 77 in 2021 without winning the state.

Did the Bharatiya Janata Party build its 2014 majority mainly in the Hindi belt?

Predominantly, though not exclusively. Of the party’s 282 seats in 2014, the ten Hindi belt states supplied roughly two thirds, including 71 of 80 seats in Uttar Pradesh, all 25 in Rajasthan, 27 of 29 in Madhya Pradesh and all 7 in Delhi, a concentration no previous majority had matched.

The remaining third came from Gujarat’s 26, Maharashtra’s 23 in alliance, Karnataka’s 17 and Assam’s 7, among smaller hauls, which is why the 2019 campaign targeted the east and south for the growth that took the party to 303.

The strategic literature inside the framework adds the organisational stratum. The party’s booth-level machinery, its membership drives conducted by missed telephone call, its panna pramukh system of page-level voter custodians, and the Rashtriya Swayamsevak Sangh’s deployment of pracharaks into election management are documented in the accounts of journalists and scholars including Rahul Verma and Pranav Gupta’s analyses of the party’s organisational advantage in the Economic and Political Weekly’s post-election issues of 2014 and 2019. Those Economic and Political Weekly special issues, the under-cited archive of the whole comparator debate, recorded within weeks of each verdict the constituency-level patterns that the three frameworks later theorised: the caste composition of the Bharatiya Janata Party’s 2014 vote, the state-wise divergence of its 2019 growth, and the opposition’s recovery geography in 2024. A researcher assembling interview recordings and field notes from those campaigns, and preserving them in a secure vault journalists use to protect their sources, holds primary material for the framework’s next edition, whichever scholar writes it. The comparator point stands above the detail: the ascent was real, it was organisational before it was rhetorical, and it was achieved inside a party system whose other dimensions, deepening and fragmentation, conditioned every step of it.

The Federal Dimension: Regional Parties as Governors of the System

Each framework was written primarily from national materials, and each is corrected by the same federal fact: between 1989 and 2024, the governments of India’s large states were held by regional parties, or by state units acting with near-total autonomy, for most of the period in most of the states that decide national outcomes. The comparative anatomy of those state settlements is set out in the comparative map of state political settlements that opened this series’ eighth pillar. Tamil Nadu was governed by the Dravida Munnetra Kazhagam or the All India Anna Dravida Munnetra Kazhagam in unbroken alternation. Odisha was governed by the Biju Janata Dal under Naveen Patnaik from March 2000 to June 2024, twenty four years in which the party contested national elections on a doctrine of equidistance and converted state incumbency into a parliamentary bloc that both national alliances courted. Andhra Pradesh was governed by the Telugu Desam Party under N. Chandrababu Naidu during the very years, 1996 to 2004, when his party’s parliamentary support underwrote first the United Front and then the National Democratic Alliance, a leverage that made a state leader the convener of a national coalition, in the period surveyed by the 1998 NDA’s construction under Vajpayee. West Bengal passed from the Left Front to the Trinamool Congress in 2011 without the national parties recovering a governing role in either era.

The federal dimension is not a fourth framework, and this article does not promote it into one, because it does not answer a fourth question of the same rank. It is better understood as the medium in which the three dimensions operate, the arena structure that decides how deepening, fragmentation and mobilisation combine in any given decade. Deepening arrived state by state because elections to state assemblies, where Other Backward Classes assertion first won office, were the upsurge’s original terrain, and because caste demography, the raw material of Yadav’s dimension, is state-specific: a Yadav plurality is a Bihar and Uttar Pradesh fact, a Maratha plurality a Maharashtra fact, a Vokkaliga and Lingayat balance a Karnataka fact. Fragmentation took the form it did because the Constitution’s Seventh Schedule and the fiscal system leave state governments in command of the patronage resources, police, land, water, agriculture and public health among them, that Chhibber’s parties distribute; a leader splitting a national party loses little, because the resources of linkage sit in the state capital. Mobilisation succeeded state by state because the Sangh’s organisation and the Bharatiya Janata Party’s campaigns were built provincially over decades, and because the party’s breakthroughs, Gujarat in 1995, Karnataka in 2008, Assam in 2016, Uttar Pradesh in 2017, each required a distinct provincial settlement with local caste and factional material.

The federal medium also explains the period’s two great asymmetries. The first is that national verdicts and state verdicts repeatedly diverged: the Bharatiya Janata Party won Lok Sabha majorities in 2014 and 2019 in states whose assemblies it lost within months or years, Delhi in 2015 and 2020, Bihar in 2015, West Bengal in 2021, Karnataka in 2023, and it lost its Lok Sabha majority in 2024 in states, Uttar Pradesh above all, whose assembly it had swept in 2022. The second asymmetry is that the opposition’s road back to national relevance ran through state governments: every serious challenge to the post-2014 order was incubated in a state incumbency, in Delhi, in West Bengal, in Telangana after 2023 under the Congress, in Karnataka after 2023, and the INDIA coalition of 2024 was in substance a federation of state incumbencies and state-level caste coalitions. A scholar reading only national aggregates would misdate nearly every turning point of the period by one election cycle, and the comparator reading, by locating its three dimensions inside the federal medium, insulates itself against that error. The point also completes the adjudication of scale between the frameworks: Yadav’s waves are dated by state elections as much as by parliamentary ones, Chhibber’s party systems are state systems in the first instance by his own account, and Jaffrelot’s silent revolution is explicitly a set of provincial revolutions proceeding at different speeds toward a northern completion.

The medium kept generating its characteristic events into the period’s final years, and two Maharashtra episodes belong in the record because all three frameworks must digest them. In June 2022 the Shiv Sena, a party founded in 1966 and led after 2012 by Uddhav Thackeray, split when Eknath Shinde carried a majority of its legislators into a new government with the Bharatiya Janata Party, and the Election Commission subsequently awarded the party’s name and symbol to the Shinde faction. In July 2023 the Nationalist Congress Party split on the same axis, Ajit Pawar carrying a legislative majority into the ruling alliance against his uncle Sharad Pawar, the party’s founder, with the name and symbol again awarded to the legislative majority. The episodes are Chhibber’s fragmentation in its purest laboratory form, parties dividing along legislative and familial lines with the state as the prize, executed inside a period of supposed national consolidation; they are also, for Jaffrelot’s framework, the mobilisation’s alliance method applied to Maharashtra’s parties one by one, and for Yadav’s, a reminder that the deepened electorate’s parties remain organisationally shallow in the dimension his framework does not measure. The comparator reading needs no further exhibit of the dimensions’ simultaneity.

Party Families after Mandal: Trajectories the Three Readings Must Share

A comparator reading is only as good as the trajectories it can state without help from any framework, so the party families’ records deserve a section in the indicative mood before the adjudication begins. The Congress entered the period as the system’s former centre and exited its third decade as one party among several, albeit the only one besides the Bharatiya Janata Party with a presence in nearly every state. Its Lok Sabha sequence, 197 in 1989, 244 in 1991, 140 in 1996, 141 in 1998, 114 in 1999, 145 in 2004, 206 in 2009, 44 in 2014, 52 in 2019 and 99 in 2024, describes a party that lost its northern base first, to caste parties in Uttar Pradesh and Bihar during the 1990s, retained governmental relevance through alliances in the 2000s, and then lost its claim to natural incumbency to the Bharatiya Janata Party in the 2010s. The organisational chapters of that story, the atrophy of district committees, the nomination of candidates by central leaderships, the absence of internal elections with contested outcomes across most decades, belong to this series’ seventh pillar and to the post-2014 Congress organisational decline in particular; for the party-system question, what matters is that all three frameworks can describe the trajectory, and each explains a different part of it: Yadav the social narrowness of its old coalition, Chhibber the organisational shell that remained when the coalition left, Jaffrelot the opponent whose mobilisation annexed its voters state by state.

The caste parties of the north describe the period’s most dramatic arc, an arc the frameworks value very differently. The Samajwadi Party won 36 Lok Sabha seats in 2004 and governed Uttar Pradesh in 2003 to 2007 and again, with a full majority of 224 assembly seats, in 2012 to 2017 under Akhilesh Yadav; the Bahujan Samaj Party won 19 Lok Sabha seats in 2004 and its assembly majority of 2007 has already entered this article as the upsurge’s emblem. Both parties then contracted sharply under the Bharatiya Janata Party’s post-2014 pressure: the Bahujan Samaj Party won no Lok Sabha seat in 2014 despite a fifth of Uttar Pradesh’s vote, recovered 10 seats in 2019 in alliance with its old rival, and fell to none in 2024; the Samajwadi Party, reduced to 5 seats in 2014 and 2019’s alliance arithmetic, rebounded to 37 seats in 2024 as the senior partner of the INDIA coalition in the state, the single largest state contingent of any opposition party. The Rashtriya Janata Dal’s Bihar arc runs from dominance in the 1990s through exclusion from power after 2005 to a return as the largest single party in the Bihar assembly of 2020 with 75 seats, while its rival and sometime partner, the Janata Dal (United), governed Bihar for most of the period from 2005 onward by re-engineering the caste coalition beneath it, adding Mahadalit and Extremely Backward Classes categories to the state’s political vocabulary. The lesson the comparator reading draws is temporal: caste parties were the system’s protagonists for its first twenty years and its survivors thereafter, weakened as governments, persistent as vote-holders, and in 2024 indispensable again to any national opposition arithmetic.

The Left’s trajectory is the period’s clearest decline. The Communist Party of India (Marxist) and its allies won 60 seats in 2004 and used them to condition a national government’s programme for four years; the party won 16 seats in 2009 after its West Bengal bastion began to crack, 9 in 2014 after losing Bengal in 2011, 3 in 2019 and 4 in 2024, retaining governmental power only in Kerala, where the Left Democratic Front won re-election in 2021 against the state’s four-decade habit of alternation. The decline’s causes are contested, industrial policy in Bengal, the Singur and Nandigram crises of 2006 to 2008, the national party’s subordination to its Bengal unit’s needs among them, but the party-system fact is common ground: the ideological pole that the Congress System had contained and the coalition era had empowered vacated the national stage during the very decades in which Chhibber and Verma found ideological structuring in voter choice, a tension the adjudication notes. The regional parties outside the caste family, the Telugu Desam Party, the Biju Janata Dal, the Trinamool Congress, the Dravida parties, the Nationalist Congress Party and its successors, complete the census: several of them governed their states for most of the period, two of them, in Andhra Pradesh and West Bengal, displaced entrenched incumbents and built dominance of their own, and their collective parliamentary weight decided the governments of 1996, 1998, 1999 and 2024.

Set against these trajectories, the three frameworks’ silences become informative. Yadav’s framework has the least to say about the Left’s decline, which was a decline of a party of the plebeian left during the plebeian upsurge, and its honest treatment requires the admission that the upsurge’s vehicles were caste parties, not class parties, a distinction Jaffrelot’s caste tables had already made unavoidable. Chhibber’s framework has the least to say about why the Bharatiya Janata Party, among all the fragments, consolidated: his measures register the consolidation without explaining the mobilisational labour behind it. Jaffrelot’s framework has the least to say about the southern systems, where his dialectic of caste assertion and Hindu consolidation barely engaged Tamil Nadu or Kerala across the whole period. A total theory would have to fail somewhere; the comparator reading prefers three theories that know their borders.

Three Dimensions, One System: The Comparator Map

The comparator reading can be stated as a table, and the table is this article’s findable artifact: a three-framework comparator map in which each framework is fixed by the question it answers, the dimension it measures, the evidence it privileges and the verdict it reaches on the period. Readers who cite one row against another mistake the artifact’s purpose; the rows are coordinates, not candidates.

Table 1, the three-framework comparator map, sets Yogendra Yadav’s three-wave framework, Pradeep Chhibber’s fragmentation framework and Christophe Jaffrelot’s Hindu-nationalist-mobilisation framework side by side on the same post-Mandal record.

Framework	Analytical dimension	Governing question	Primary evidence	Verdict on the post-Mandal system
Yadav, three waves of democratic deepening (essays from 1996; Making Sense of Indian Democracy, 2020)	Substantive democratic content	What changed for citizens placed lowest in the caste order?	Lokniti National Election Studies turnout and representation by caste; caste profile of legislators	A third wave of deepening: lower-caste participation and representation reached the northern heartland after 1990
Chhibber, fragmentation (Democracy without Associations, 1999; Ideology and Identity with Rahul Verma, 2018)	Institutional architecture of competition	How is competition organised after the Congress System’s end?	Effective number of parties; state party systems; party organisation and splits; government duration	A fragmented architecture, later inhabited by ideological competition on statism and recognition
Jaffrelot, Hindu nationalist mobilisation (The Hindu Nationalist Movement and Indian Politics, 1996; India’s Silent Revolution, 2003; Modi’s India, 2021)	Political strategy and regime character	How did the period’s most successful movement convert assertion into power?	Legislator prosopography by caste; Sangh organisational record; state campaign histories	A silent revolution of lower-caste power, answered by a majoritarian mobilisation that won the system after 2014

The map disciplines the comparator claim in two ways. First, it shows that the frameworks’ central terms do not translate into one another. Deepening is not the opposite of fragmentation: a system can deepen and fragment simultaneously, and the Indian system after 1990 did both, admitting new social groups to power through a multiplying set of parties. Fragmentation is not the opposite of mobilisation: the Bharatiya Janata Party’s ascent was executed through the fragmented architecture, state by state and ally by ally, before it was executed against it. Mobilisation is not the opposite of deepening: the ascent depended on the deepened electorate, recruited the upsurge’s castes and cast its leader as the upsurge’s product. Second, the map shows what each framework would count as its own falsifier, which totalising readings rarely specify. Yadav’s row falls if lower-caste participation and representation are shown to have stagnated after 1990; they did not. Chhibber’s row falls if national linkages are shown to have remained intact after 1989; they did not. Jaffrelot’s row falls if the Bharatiya Janata Party’s rise is shown to owe nothing to caste arithmetic and movement organisation; it owed a great deal to both. Three unfalsified answers to three different questions is the comparator reading’s whole case, and it is enough.

Can fragmentation and a single-party majority describe the same system at the same time?

Yes, because the two terms measure different properties. Fragmentation, in Chhibber’s usage, describes how competition is organised: the number and state-anchorage of parties and the divisibility of their linkages to voters. A majority describes one election’s seat outcome under first-past-the-post rules, which in 2014 converted 31.3 percent of the vote into 282 seats.

The organised fragmentation persisted beneath the outcome, visible in the state results of 2015 to 2023, and it re-emerged in the Lok Sabha itself in 2024, when 240 seats left the largest party dependent on allies whose bases lie in single states.

The map also locates the pillar’s other instruments. The decisive-moment reading of Mandal, built at the hub, asked how a single decision reorganised the system’s possibilities; the comparator map assumes that reorganisation and asks what kind of system resulted. The state-settlement comparisons of the eighth pillar supply the map’s missing fourth coordinate, the federal one: each framework’s dimension varies by state, and the all-India row for any framework is an average of Tamil Nadu’s early deepening, Bihar’s upsurge and Gujarat’s consolidation. Scholars who extend the comparator project will likely add that federal coordinate as a dimension in its own right rather than as a footnote, and the state evidence assembled across this series, from Uttar Pradesh as bellwether and laboratory to the Kerala alternation studies, gives them the material. The comparator reading claims completeness only for its three dimensions, not for the space of all possible dimensions, and that modesty is part of its method.

One further use of the map deserves note, because it disciplines citation beyond this series. Debates about Indian politics in newspapers and classrooms repeatedly import a conclusion earned in one dimension as if it were earned in all three: a finding about turnout is quoted against a claim about party organisation, a finding about one party’s strategy is quoted against a claim about the electorate’s depth. The map gives editors and teachers a mechanical check against that importation. Ask of any striking claim about the post-Mandal system which row’s evidence could possibly establish it, and a large share of the period’s polemics dissolves before the evidence is even consulted.

The Shared Archive: Surveys, Special Issues and the Making of the Evidence

A comparator reading is possible at all because the three frameworks drink from the same wells, and the wells deserve a section, since their construction shaped what the frameworks could see. The first well is the Lokniti programme at the Centre for the Study of Developing Societies, which conducted the National Election Studies of 1996, 1998, 1999, 2004, 2009, 2014, 2019 and 2024 with samples in the tens of thousands, recording each respondent’s caste, sub-caste, class, religion, locality and vote. Before Lokniti, claims about caste and voting rested on constituency ecology and journalistic reportage; after Lokniti, they rested on survey marginals that any scholar could recompute, and the frameworks’ shared dependence on those marginals is the strongest guarantee that their disputes are interpretive rather than evidentiary. Yadav directed the programme’s early waves, Chhibber and Verma drew on its data for the attitudinal batteries of Ideology and Identity, and Jaffrelot, whose own archive is the legislator census, cites its post-election estimates for the caste composition of the Bharatiya Janata Party’s vote in 2014 and 2019. The survey tradition also produced the period’s standard turnout series by social group, the raw material of the upsurge thesis, refined in the State of Democracy in South Asia studies and in the work of Sanjay Kumar, the programme’s subsequent director.

The second well is the Economic and Political Weekly’s post-election special issues, the under-cited primary archive named in this article’s brief. Within weeks of the 2014 verdict, the journal’s special issue carried constituency and survey analyses by Suhas Palshikar, K. K. Kailash, E. Sridharan and the Lokniti team that fixed the first draft of every later interpretation: the geographic concentration of the Bharatiya Janata Party’s majority, the caste profile of its vote, the Congress’s collapse among its own former constituencies. The 2019 special issue repeated the exercise with Rahul Verma and Pranav Gupta’s account of the ruling party’s organisational machine and Palshikar’s mapping of the verdict’s state-wise structure. The 2024 special issue, appearing after a verdict no framework had forecast in detail, recorded the opposition’s recovery geography in Uttar Pradesh and Maharashtra and reopened the question of the system’s direction at the moment the frameworks were most confident. A scholar who reads the three special issues in sequence watches the comparator debate being born three times, each time on fresh tables, and watches each framework select its rows.

The third well is the official record itself, the Election Commission’s statistical reports, supplemented from 2004 by the affidavit archive created when the Supreme Court’s judgments of 2002 and 2003 compelled candidates to declare assets, liabilities and pending criminal cases, declarations compiled and published by the Association for Democratic Reforms. That archive changed the period’s scholarship in a direction none of the three frameworks anticipated, a point the next section’s discussion of money and criminalisation takes up. Here it completes the comparator’s evidentiary picture: the frameworks disagree inside an unusually well-lit room, and the lighting, official counts, mass surveys and sworn affidavits, is itself a product of the democratic order whose character they dispute. Yadav has made the point in his own way, that a democracy which studies itself this closely has already deepened in one dimension whatever the other dimensions show.

Where the Frameworks Collide: Three Pairwise Tensions Adjudicated

Complementarity has been the claim; collision is the test. The Yadav and Chhibber frameworks collide over the meaning of the same multiplication of parties. For Yadav, the Samajwadi Party’s founding in 1992, the Bahujan Samaj Party’s growth and the Janata Dal’s state successors are the institutional form of democratisation: groups that had voted for other people’s parties acquired parties of their own, and the fragmentation of the 1990s is the sound of the caste order’s lower floors entering the building. For Chhibber, the same events register as the failure of association and aggregation: parties multiplied because nothing held them together, programmes thinned into leader appeal and bloc arithmetic, and the system’s capacity to convert votes into accountable government weakened. The adjudication turns on a distinction both scholars can accept: entry and aggregation are different functions, and a system can improve at the first while deteriorating at the second. The 1990s record shows both movements at once. Entry indisputably democratised, measured by the caste of candidates, legislators and chief ministers. Aggregation indisputably strained, measured by the five elections and nine governments of the decade’s first half and by the outside-support improvisations of 1996 to 1998. The tension dissolves into the dimensions, but a residue remains, and honesty requires naming it: Yadav’s framework carries a normative warmth toward the multiplication that Chhibber’s lacks, and Chhibber’s carries an institutional pessimism that Yadav’s lacks, and those tempers are not evidence. The comparator reading keeps both tempers visible and trusts neither.

The Chhibber and Jaffrelot frameworks collide over 2014. For a fragmentation reading stated carelessly, the Bharatiya Janata Party’s majority refutes the period’s defining claim, and some of Chhibber’s critics stated his claim carelessly in order to refute it. Stated precisely, as the architecture of competition rather than as a forecast of seat shares, the framework survives 2014 and is vindicated in an unexpected quarter: the majority was assembled through the fragmented system’s own instruments, pre-electoral state alliances, state-specific campaigns, caste-wise candidate selection and a national leadership tour that functioned as a federation of state campaigns. Jaffrelot’s framework explains why this party, and not another fragment, performed the assembly: only the Sangh’s cadre organisation and the temple movement’s accumulated constituency gave a party both the ground machinery and the ideological cement that aggregation required. The collision’s residue here is a question of causal depth. Chhibber’s dimension explains the opportunity structure of 2014, the absence of any rival aggregator; Jaffrelot’s explains the agent who exploited it. Neither explanation contradicts the other, but they compete for the title of the deeper cause, and the comparator reading awards that title to neither, because depth in causal accounts of elections is a property of the question asked, and the questions differ. What the record does settle is sequencing: the architecture fragmented first, between 1989 and 1996, and the mobilisation that eventually mastered it required eighteen further years, two defeats and a change of leadership generation, which is a long lag for a supposedly inevitable consolidation.

The Yadav and Jaffrelot frameworks collide over the character of the post-2014 order, and this collision is the deepest of the three because it is partly about facts of valuation rather than of count. Yadav reads the Bharatiya Janata Party’s lower-caste coalition, its Other Backward Classes prime ministerial candidate in 2014, its recruitment of Extremely Backward Classes and non-dominant Scheduled Castes, and its welfare delivery pitched to named beneficiary categories, as the third wave working through the system’s winner: the upsurge set the terms, and even a Hindu nationalist government must rule as a coalition of the deepened electorate. Jaffrelot reads the same facts within a strategy of containment: caste is courted electorally in order to be subordinated politically to Hindu unity, the communities mobilised are the smaller and less organised ones played against the dominant backward castes, and the regime’s institutional conduct toward minorities marks the order as majoritarian in principle whatever its electoral sociology. The electoral record adjudicates the descriptive half of this dispute and leaves the evaluative half open, and the comparator reading says which half is which. Descriptively, both are right about their own rows: the 2014 and 2019 winning coalitions were the most caste-inclusive in the Bharatiya Janata Party’s history, on Lokniti’s post-election estimates, and the same governments pursued a legislative and administrative programme, on citizenship, on Kashmir’s status and on religious conversion and cow protection in the states, whose majoritarian character is documented and disputed in the terms the stance charter of this series requires to be attributed. Evaluatively, whether inclusion as voters outweighs subordination as castes is a judgement about what democracy is for, and frameworks that answer it differently are not failing a test the record can set. The comparator reading’s contribution is to stop the two scholars’ readers from treating that evaluative dispute as if it were a dispute about turnout.

The general election of 2024 supplied the collision’s freshest evidence, and it cut in both directions at once, which is the comparator pattern in miniature. In Uttar Pradesh, the Samajwadi Party campaigned on an explicitly upsurge-era formula, the PDA coalition of the backward, Dalit and minority communities, fielded candidates across the caste order’s lower two thirds, and won 37 seats against the Bharatiya Janata Party’s 33 in the state, including Faizabad, where the temple consecrated in January 2024 stands. Yadav’s readers claimed the result as the third wave’s resilience: the deepened electorate, offered a vehicle of its own, withdrew its loan to the mobilisation. In Bihar, by contrast, the Janata Dal (United) and the Bharatiya Janata Party divided the state’s seats between them on an arithmetic of Extremely Backward Classes and Mahadalit engineering that both parties had borrowed from the upsurge’s own toolkit, and the Rashtriya Janata Dal’s campaign, led by Tejashwi Yadav on jobs and caste census, won only 4 of the state’s 40 seats. Jaffrelot’s readers claimed Bihar as the accommodation’s maturity: the mobilisation had learned to win the upsurge’s castes on the upsurge’s terrain. Both claims are documented; neither cancels the other; and the two states together demonstrate why a framework war fought on national aggregates alone would have declared a victor the evidence does not contain. The comparator reading’s insistence on state-by-state adjudication, election by election, is not pedantry. It is the only procedure under which all three frameworks can be true at once in the places where each is strongest.

The Strongest Objection: Disagreement about Facts, Not Only Dimensions

The strongest counter to the comparator reading, flagged in the brief for this article and pressed by the frameworks’ partisans, is that the complementarity is an artifact of charitable paraphrase: state the frameworks precisely, the objection runs, and they do contradict one another about observable matters. Three alleged contradictions deserve examination rather than dismissal. First, Yadav and Chhibber disagree about the post-1990 voter: Yadav’s voter is increasingly autonomous, voting caste as an assertion of selfhood; Chhibber and Verma’s voter of 2018 is increasingly ideological, sorting by statism and recognition. These are rival accounts of the same Lokniti datasets, and they cannot both be wholly right. The comparator verdict is that they are answers about different strata of the same act: a voter can assert identity in choosing which parties are imaginable and sort ideologically among the parties imagined, and the survey instruments that detect the first movement, caste-wise turnout and vote shares, are not the instruments that detect the second, attitude batteries on state intervention and minority protection. The contradiction is real at the level of emphasis, and it is the kind of dispute that further survey waves can narrow; it is not a contradiction about the seat record, and it does not touch the dimensional division of labour this article defends.

Second, Jaffrelot’s ethnic democracy characterisation and Yadav’s deepening characterisation appear to contradict one another about the regime itself. The comparator verdict distinguishes the time horizons their evidence commands. Yadav’s data series run from 1952 to the present and measure the social depth of participation; Jaffrelot’s characterisation addresses a decade’s institutional conduct. A system can be, simultaneously, the deepest democracy in its social history and a regime whose ruling formation asserts majority primacy; indeed the second condition may require the first, since majoritarian consolidation in a universal-franchise system needs the majority community’s lower orders inside the coalition, which is precisely the upsurge’s bequest. The comparator reading therefore treats the apparent contradiction as the period’s central paradox rather than as an error in either framework, and it predicts that scholarship of the next decade will be organised around that paradox, as the Economic and Political Weekly’s 2024 post-election issue already began to be, with its contributors dividing over whether 2024 represented the system’s self-correction or its stress fracture.

Third, Chhibber’s fragmentation and the very idea of a post-Mandal system appear to contradict one another: a fragmented field, the objection runs, is not a system. The verdict here is definitional and short. A party system, in the comparative literature from Sartori onward, is a pattern of interactions among parties competing for office, and patterned fragmentation is a pattern: India’s post-1990 interactions, alliance conventions, state duopolies summed into national pluralities, and the regular translation of 31 to 37 percent vote shares into parliamentary majorities or near-majorities, are as patterned as the Congress dominance they replaced. The objection fails, and its failure strengthens the comparator architecture, because it shows that even the period’s name survives the fragmentation framework’s own scrutiny.

A further delimitation concerns what all three frameworks omit, because a comparator reading that only rearranged its three sources would inherit their common blind spot. The period’s scholarship outside the three frameworks documents a transformation none of them places at its centre: the rising cost and criminalisation of candidacy. The Association for Democratic Reforms’ analyses of sworn affidavits record the share of Lok Sabha members facing pending criminal cases rising from roughly 30 percent in 2009 to 43 percent in 2019 and 46 percent in 2024, and Milan Vaishnav’s When Crime Pays, published in 2017, explains the pattern functionally, as the rational choice of parties needing self-financing candidates in a system where elections are ruinously expensive and where voters in weakly governed districts read a strongman’s record as a credential of protection. Campaign finance moved in parallel, through the electoral bond scheme of 2018, struck down by the Supreme Court in February 2024, whose published data showed the ruling party redeeming the majority of bonds sold. None of the three frameworks denies this record; none of their dimensions, democratic content, institutional architecture or mobilisation strategy, was designed to hold it. The comparator reading therefore records money and criminalisation as the strongest candidate for a fourth dimension, one a successor synthesis will have to build, and notes that its absence from the comparator map marks the map’s date as much as its design.

What remains after the objection has been given its due is a delimitation of the comparator claim, which this article accepts and states as its final analytical position. The three frameworks are complementary in the dimensions they measure and incomplete in the dimensions they omit, they collide in emphasis and in valuation more than in fact, and the collisions that are factual, about voter motivation above all, are empirical programmes for the Lokniti tradition all three scholars have used. The claim first advanced in this article’s opening stands, with its scope specified: the post-Mandal party system invites three complementary analytical characterisations, and the comparator-framework reading captures their complementary scopes.

What the Comparator Reading Settles and What It Leaves Open

A series that argues must end by separating its settlements from its open questions. Settled, on the record assembled here: the post-1990 decades transformed the social composition of Indian political power, and no return to the pre-1990 composition occurred in any subsequent election, including those of 2014, 2019 and 2024. Settled: the architecture of competition fragmented after 1989 and has not been rebuilt into the integrated national form the Congress System embodied, the majority decade of 2014 to 2024 having been achieved through the fragmented architecture and having ended, in 2024, inside it. Settled: the Bharatiya Janata Party’s ascent was the period’s largest strategic fact, it was built by movement organisation and caste arithmetic over thirty five years, and it reorganised every other party’s choices, from the Congress’s alliance strategy to the caste parties’ survival strategies to the Left’s marginalisation. Settled, finally: the three frameworks that name these settlements, Yadav’s, Chhibber’s and Jaffrelot’s, answer different questions, and the comparative literature’s habit of staging them as a knockout tournament has cost more understanding than it has produced.

Open, and left open deliberately: whether the deepening Yadav chronicled can resume its advance from representation into social and economic equality, or whether the upsurge’s parties have exhausted their programme at the threshold of power. Open: whether the architecture Chhibber mapped is entering a phase of re-aggregation around two national poles with state franchises, a pattern some scholars of the 2024 verdict discern in the INDIA coalition’s consolidation of the opposition vote, or whether 2024 was a fluctuation inside durable fragmentation. Open: whether the regime character Jaffrelot named will be consolidated, moderated by coalition dependence, or reversed by electoral alternation, a question the state results of 2015 to 2024 answer in contradictory directions and the next general election will answer again. Each framework also issues a testable expectation for the decade that follows the period, and recording the expectations is the comparator reading’s final service, because it converts a synthesis into a wager that later evidence can settle. Yadav’s dimension predicts that lower-caste participation and representation will not recede under any governing formation, and that parties will compete to extend the upsurge’s frontier, to Extremely Backward Classes within the backward bloc, to sub-castes within the Scheduled Castes, to women as a constituency addressed in their own name; the sub-categorisation litigation and the caste census demands taken up in this pillar’s closing articles are that prediction already in motion. Chhibber’s dimension predicts that no governing majority will be secure unless it is assembled as a federation of state settlements, and that national parties will continue to be punished in states where their state units are shells; the verdict of 2024, in which the ruling party’s losses concentrated in its organisationally hollowest large state, is the prediction’s first confirmation. Jaffrelot’s dimension predicts that the Hindu nationalist formation will answer electoral setbacks by deepening, not diluting, its twin registers, cadre organisation below and majoritarian legislation above, and that the tension between its caste arithmetic and its ideological programme will define the opposition’s opportunities. The three predictions do not contradict one another, which the comparator reading counts as its closing evidence.

The pillar’s remaining articles take up the sub-categorisation question, the economically weaker sections amendment and the caste census debate, each of which will redistribute the evidence among the three dimensions; readers who hold the comparator map loosely, as a map and not as a verdict, will be best placed to see the redistribution when it comes. The post-Mandal party system has outlasted the system that preceded it, and the scholarship that reads it in three dimensions at once reads it whole.

Frequently Asked Questions

Q: What is the post-Mandal party system?

The post-Mandal party system is the pattern of party competition that followed the implementation, in August 1990, of the Mandal Commission’s recommendation of reservation for Other Backward Classes in central government employment. Its marks are the collapse of Congress dominance after 1989, the rise of caste-based and regional parties in the states, a quarter century without a single-party Lok Sabha majority between 1989 and 2014, and the eventual consolidation of the Bharatiya Janata Party, which won majorities in 2014 and 2019 before returning to coalition dependence in 2024. Scholars read the system through three frameworks, associated with Yogendra Yadav, Pradeep Chhibber and Christophe Jaffrelot, which this article treats as complementary.

Q: Who is Yogendra Yadav?

Yogendra Yadav is an Indian political scientist and psephologist, long associated with the Centre for the Study of Developing Societies in Delhi and its Lokniti programme of election studies, which he helped build into the principal survey record of Indian voting behaviour. His scholarship on the post-1990 decades produced the three-wave framework of democratic deepening and the concept of the second democratic upsurge, describing the surge of lower-caste participation in northern politics. His essays are collected in Making Sense of Indian Democracy, published in 2020. He later entered electoral politics himself as a founder of Swaraj India, an unusual trajectory for the period’s leading analyst of party competition.

Q: Who is Pradeep Chhibber?

Pradeep Chhibber is a political scientist at the University of California, Berkeley, whose work studies how party systems organise competition. His Democracy without Associations, published in 1999, argued that Indian parties developed without dense civic associations linking them to society, and connected themselves to voters through state patronage instead, producing a fragmented party system after the Congress System ended. With Rahul Verma, he wrote Ideology and Identity in 2018, arguing from survey evidence that Indian competition is structured by two ideological dimensions, statism and recognition. His framework supplies the institutional-architectural reading of the post-Mandal system.

Q: Who is Christophe Jaffrelot?

Christophe Jaffrelot is a French political scientist at the Centre for International Studies and Research in Paris and at King’s College London, and the leading historian of Hindu nationalism among the period’s scholars. His The Hindu Nationalist Movement and Indian Politics appeared in 1996, and India’s Silent Revolution, published in 2003, documented the rise of the lower castes in north Indian politics through seven decades of legislator data. Modi’s India, published in 2021, extended his account to the post-2014 order, which he characterises as an ethnic democracy. His framework supplies the political-strategic reading of the post-Mandal system, centred on mobilisation and its consequences.

Q: What is the three-wave democracy framework?

The three-wave democracy framework is Yogendra Yadav’s periodisation of Indian democratic deepening. The first wave, from 1947, established formal democracy: universal franchise, an election commission, courts and legislatures. The second wave, from the late 1960s through the 1980s, widened participation from inside the old order, as Other Backward Classes leaderships rose in the southern and western states and opposition parties proved incumbents could be removed. The third wave, from 1989 and 1990, carried the widening into the northern heartland and the lower reaches of the caste order, with lower-caste citizens voting, winning and governing in proportions approaching their population shares.

Q: What is institutional-architectural fragmentation?

Institutional-architectural fragmentation is Pradeep Chhibber’s characterisation of the party system that followed the Congress System. It describes a condition in which competition is organised through many parties anchored in single states, linkages between parties and voters run through patronage rather than associations, leadership successions are personal and often familial, and the national legislature aggregates separate state contests instead of hosting one national contest. Its measures include the effective number of parties, which rose above five in the 1990s, and the share of seats held by state-confined parties. The term names the system’s architecture, not its election outcomes, so it can coexist with a single-party majority.

Q: How do the three frameworks of post-Mandal politics relate?

The three frameworks relate as dimensions of one system rather than as rival totals. Yogendra Yadav’s three-wave framework measures substantive democratic content: what changed for lower-caste citizens. Pradeep Chhibber’s fragmentation framework measures institutional architecture: how competition came to be organised after Congress dominance ended. Christophe Jaffrelot’s Hindu-nationalist-mobilisation framework measures political strategy: how the period’s most successful movement converted caste assertion into national power. Their evidence overlaps, since all three use Election Commission records and Lokniti surveys, and their disputes concern which dimension is primary and how the post-2014 order should be valued, matters this article adjudicates as questions of emphasis and valuation more than of fact.

Q: What does Chhibber mean by democracy without associations?

Democracy without Associations, published in 1999, is Pradeep Chhibber’s diagnosis of how Indian parties connect to society. In the classic European cases, parties grew inside dense networks of trade unions, churches, professional bodies and civic associations that aggregated interests, recruited members and enforced programmes. Chhibber argues that Indian parties developed without such a network, and therefore linked themselves to voters through the state: patronage, discretionary welfare, public employment and appeals to enumerated caste blocs. The consequence is a party system prone to fragmentation, because linkages built on divisible benefits can be replicated by any leader who breaks away with a bloc, a district machine or a ministry’s resources.

Q: What is the second democratic upsurge?

The second democratic upsurge is Yogendra Yadav’s name, in a 2000 essay, for the surge of lower-caste political participation in northern India during the 1990s. A first upsurge had occurred in the 1960s and 1970s, when Other Backward Classes entered electoral politics in force in the southern and western states. The second upsurge carried the same movement into Uttar Pradesh, Bihar and Madhya Pradesh after 1990: turnout among Other Backward Classes and Scheduled Castes matched or passed upper-caste turnout in survey estimates, lower-caste legislators multiplied, and parties led by the mobilised castes themselves, the Samajwadi Party, the Bahujan Samaj Party and the Janata Dal’s successors, took power in the northern states.

Q: How did OBC representation in the Lok Sabha change between 1989 and 2004?

Christophe Jaffrelot’s tabulations of Election Commission candidate records show a steep transfer of seats in the Hindi belt. Among members elected from the northern states, the Other Backward Classes share rose from roughly one in nine in 1984 to roughly one in four by 1996, and held near that level into the 2000s, while the upper-caste share fell from around one half in the early decades to under one third. The change was driven by caste parties fielding their own candidates, by the Janata Dal’s and later the Bharatiya Janata Party’s recruitment, and by the Mandal implementation’s legitimation of backward-caste claims to office.

Q: What is the effective number of parties, and what did it show in the 1990s?

The effective number of parties is a standard comparative measure that counts parties weighted by their shares of votes or seats, so that a legislature of many tiny parties and one large party scores lower than a legislature of several equal parties. Applied to India, the measure rose from under three in the Lok Sabha elections of the early 1980s to above five during the 1990s, reaching its height in the elections of 1996 to 1999, when more than thirty parties won seats and the largest held under a third of them. Pradeep Chhibber’s fragmentation framework rests heavily on this series, which fell back after 2014 without returning to the levels of the Congress dominance decades.

Q: Why did caste-based parties lose ground in the 2014 and 2019 elections?

Several pressures converged. The Bharatiya Janata Party built its own caste coalition, recruiting non-Yadav Other Backward Classes and non-Jatav Scheduled Castes directly and projecting a prime ministerial candidate of backward-caste origin, which cut into the caste parties’ claim to monopolise lower-caste representation. Welfare delivery pitched to named beneficiary households weakened the broker role through which caste parties had mediated the state. The caste parties also carried incumbency records, family succession disputes and, in Uttar Pradesh in 2019, an alliance between old rivals that their own voters only partly followed. Their vote shares fell less than their seats, since first-past-the-post rules magnify a divided opposition’s losses, and in 2024 the Samajwadi Party’s rebound to 37 seats showed the blocs had not dissolved.

Q: How did the BJP assemble a cross-caste coalition after 2014?

The Bharatiya Janata Party combined its upper-caste base with systematic recruitment below it. It fielded candidates from non-dominant Other Backward Classes communities, the Kurmis, Lodhs, Nishads and others whom the Yadav-dominated parties had kept from office, and from non-Jatav Scheduled Castes in Uttar Pradesh and Bihar. It allied with caste parties of those communities, including the Apna Dal and the Nishad Party, and gave their leaders ministries. It pitched welfare schemes, cooking gas connections, toilets, housing and free grain, to beneficiary households across caste lines, and it foregrounded the backward-caste origin of Narendra Modi himself. Christophe Jaffrelot’s framework reads this as the accommodation register of Hindu nationalist mobilisation; Yogendra Yadav’s reads it as the third wave dictating terms to its apparent victor.

Q: Why did the BJP need coalition partners after the 2024 election?

The Bharatiya Janata Party won 240 Lok Sabha seats in 2024, thirty two short of the majority mark of 272, because its coalition of castes and regions contracted in decisive states. In Uttar Pradesh the Samajwadi Party and the Congress together outpolled it among Other Backward Classes, Scheduled Caste and Muslim voters, and it lost seats including Faizabad, the constituency of the Ayodhya temple. In Maharashtra its alliance split the Maratha and backward-caste vote unfavourably, and in West Bengal the Trinamool Congress held its ground. The party therefore governed through the National Democratic Alliance, depending principally on the Telugu Desam Party’s 16 seats and the Janata Dal (United)’s 12, restoring the coalition dependence that had defined the system between 1989 and 2014.

Q: What happened to the Left parties after 2009?

The Left’s national position collapsed within a decade. The Communist Party of India (Marxist) won 43 Lok Sabha seats in 2004 and, with its allies, conditioned the first United Progressive Alliance government’s programme before withdrawing support in 2008. It fell to 16 seats in 2009, as its West Bengal government, in office since 1977, was damaged by the Singur and Nandigram crises, and to 9 in 2014 after Mamata Banerjee’s Trinamool Congress took the state in 2011. It won 3 seats in 2019 and 4 in 2024. The Left retained power only in Kerala, where the Left Democratic Front won re-election in 2021, and remained a significant force in Tripura’s opposition, but its role as a national coalition-maker ended.

Q: How did the national party system become a sum of state party systems?

The change followed the collapse of the Congress, the last party organised as a single national machine. From the 1990s, most large states settled into their own competitive patterns: Tamil Nadu between the two Dravida fronts, Kerala between two alliances, Uttar Pradesh among caste parties and then a national challenger, Bihar around two rival caste coalitions, West Bengal through Left dominance and then Trinamool dominance. National elections became aggregations of these state contests, fought with state-specific alliances, campaigns and caste arithmetic, and national parties learned to behave as federations of state units. Pradeep Chhibber’s collaborative work on party aggregation documented the pattern, and the Election Commission’s state-wise results for every election from 1996 onward display it directly.

Q: How did women’s turnout change across the post-1990 decades?

Women’s turnout rose steadily toward and then past parity. The Election Commission recorded a gap of about seventeen points between men’s and women’s turnout in 1962. The gap narrowed through the 1990s and 2000s as registration drives, panchayat reservation for women from 1993 onward, and party mobilisation brought women voters to the booth in growing proportions. By 2014 the gap was about one and a half points, and in 2019 women’s turnout, at 67.18 percent, exceeded men’s, at 67.01 percent, for the first time in a national election. Parties responded by pitching welfare schemes directly to women voters, a constituency Yogendra Yadav’s deepening framework counts among the third wave’s quieter achievements.

Q: How did the Samajwadi Party and the Bahujan Samaj Party differ in their caste coalitions?

The two Uttar Pradesh parties were built on different castes and different theories. The Bahujan Samaj Party, founded by Kanshi Ram in 1984, rested on the Scheduled Castes, above all the Jatavs, and sought a bahujan majority of the oppressed; under Mayawati it later courted Brahmin voters under a sarvajan appeal that won it a majority in 2007. The Samajwadi Party, founded by Mulayam Singh Yadav in 1992, rested on the Yadav and other backward castes in alliance with Muslim voters, the combination its supporters called MY. The first party began from the bottom of the caste order and bargained outward; the second began from the dominant backward caste of the state’s countryside and bargained downward and sideways.

Q: How far did the BJP’s 2019 expansion reach beyond the Hindi belt?

The 2019 verdict extended the Bharatiya Janata Party’s map in every direction except the far south. The party won 18 of West Bengal’s 42 seats, up from 2 in 2014, swept Odisha’s coastal belt with 8 seats, held all 26 in Gujarat and 23 of Maharashtra’s seats in alliance, won 25 of Karnataka’s 28, and took 4 seats in Telangana and a first foothold across the North East through its regional alliances. Its 303 seats were still weighted toward the north and west, and it won nothing in Kerala, Tamil Nadu or Andhra Pradesh, but the expansion changed the system’s geography: the principal opposition to it in 2019 was no longer a national rival but a set of state governments.

Q: Why do scholars date the post-Mandal system from 1989 rather than from 1990?

The dating dispute records the frameworks’ different breaks. The implementation decision that names the system was announced in August 1990 by the V. P. Singh government, and Yogendra Yadav’s deepening account treats that announcement as the trigger of the northern upsurge. Pradeep Chhibber’s institutional account dates the break a year earlier, to the 1989 election, when Congress dominance ended and no party won a Lok Sabha majority, because the architecture changed before the policy did. Christophe Jaffrelot’s account uses 1990 for a different reason again: the Rath Yatra of September and October 1990 launched the counter-mobilisation his framework tracks. The pillar retains the 1989 start for the system and the 1990 decision for its name.