SPECIAL ARTICLEEconomic & Political Weekly EPW february 7, 200955Principal State Level Contests and Derivative National Choices: Electoral Trends in 2004-09Yogendra Yadav, Suhas PalshikarPolitical choices in a national election increasingly derive from the competitive format, electoral cycle, political agenda, participatory pattern and social cleavages defined in state politics. In this sense, the political choices made at the state level are mostly “principal” and those made at the national level are increasingly “derivative”. But state level politics shapes and filters rather than pre-determines the national outcome. Using this framework and the trends in 2004-09, this study attempts to understand the structure of contestation that will shape the finaloutcome in the coming Lok Sabha elections. The complex pattern of principal outcomes and timing in the political calendar shows that neither of the two major national alliances can sweep the polls nor be swept away in the 2009 Lok Sabha elections. In all probability, we are going to witness one more election in which no single party or pre-poll alliance is likely to get a clear majority and one where the smallest of changes in individual states is likely to have a major impact on government formation in New Delhi.This article sets the context for the analysis of the 2008 state elections that is presented elsewhere in this issue.The authors are grateful to colleagues in the Lokniti network for an ongoing conversation on the empirical and theoretical issues raised in this essay and to the Centre for the Study of Developing Societies, Data Unit, New Delhi, in general and Himanshu Bhattacharya and Kanchan Malhotra in particular for their help with the tables used in this essay.Yogendra Yadav (yogendra.yadav2@gmail.com) is at the Centre for the Study of Developing Societies, New Delhi, and Suhas Palshikar (suhas@unipune.ernet.in) is with the Department of Politics, University of Pune, Pune.The frequent but inaccurate use of the metaphor of “semi-final” to describe the state assembly elections held in November-December 2008 draws attention to the signifi-cance of the linkage between the election verdict at the state and national levels. This metaphor recognises something that all of us have begun to take for granted in the last two decades: a transfer of political choices from the state to the national political arena. Yet the image of successively higher levels of elections leading up to a “final” that will decide the real national winner is profoundly misleading, as the Bharatiya Janata Party (BJP) discovered five years ago when it took the round of state assembly elections pre-ceding the Lok Sabha elections of 2004 as a sign of things to come. We are all wiser now. We know that temporal proximity of the recent round of elections to the Lok Sabha poll will ensure its strong influence in the six states that went to the polls–Chhattis-garh, Madhya Pradesh, Delhi, Rajasthan, MizoramandJammuand Kashmir. But there is no reason to expect that this last cluster in a series of state elections will have a special bearing on political choices made in the rest of the country in the coming Lok Sabha elections. If we wish to understand the nature of political choices in the national election, we need to understand not just the latest round of state assembly elections but the entire series of state as-sembly elections that have taken place since the last Lok Sabha elections in 2004. This is what the present essay seeks to do. This overview does not offer a narrative of the major rounds of assembly elections since 2004. Some of the articles in this issue of EPW and other writings offer the narratives and state-specific analyses.1 Instead this paper takes up two general questions: what exactly is the nature of the linkage between popular pre-ferences in state assembly elections and subsequent Lok Sabha elections? How does this affect the nature of political choices and the quality of democracy? Both these questions are addressed by reviewing electoral trends between 2004 and 2009, the period between the previous Lok Sabha elections and the coming one. The argument is presented in three parts. The first section proposes a model of linkages between state and national politics in India and distinguishes it from a similar relationship in multi-level elections in Europe. It argues that the model proposed here is specific to the context of what has been called the “third electoral system” in India. The second section is an attempt to understand all state assembly elections since 2004 by placing them in a typology. We then return to the main question of the impact of state level electoral cycles on the national verdict. This is mapped in the third section by identifying the various kinds of effects that have characterised the last two Lok Sabha elections. The implications

SPECIAL ARTICLEfebruary 7, 2009 EPW Economic & Political Weekly56of this analysis for the broader trends in Indian democracy are stated towards the end. To anticipate, the essay argues that the linkage between state level and national level choices in the last two decades has pro-duced a disjunction between the level at which political choices are made and the level at which governance decisions are effec-tively taken. This, we believe, blunts the edge of the democratic upsurge2 of the 1990s because the energy generated is directed at a political arena that is not in a position to shape public policies. Thus, while the linkage between state level and national level choices helps us understand the trend likely to emerge in the coming Lok Sabha elections, this same exercise alerts us to the routinisation of the third electoral system.3 1 Indian Model of Multiple Levels of ElectionsThe Indian experience of multiple, differentiated yet interlocked tiers of elections invites us to look for comparisons: Are there other examples of multiple levels of elections where electoral choices interweave? If yes, what is the nature of this relationship? The richest body of theoretical reflections and empirical evidence available on this question comes from attempts to understand the linkage between the elections to the European Parliament with national elections in European Union (EU) countries. Terming the direct elections to the European Parliament “second-order na-tional elections” a seminal article by Karlheinz Reif and Hermann Shmitt4 argued:The composition of the directly elected European Parliament does not precisely reflect the real balance of political forces in the European Community. As long as the national political systems decide most of what there is to be decided politically, and everything really important, European elections are additional national second-order elections. They are determined more by the domestic political cleavages than by alternatives originating in theEC, but in a different way than if nine first-order national elections took place simultaneously. This is the case because European elections occur at different stages of the na-tional political systems respective electoral cycles. Such a relationship between a second-order arena and the chief arena of a political system is not at all unusual. What is new here is that one second-order political arena is related to nine different first-order arenas. A first analysis of European election results satisfactorily justifies the assumption that European Parliament direct elections should be treated as nine simul-taneous national second-order elections. Subsequent analyses of the logic of multi-level elections in the EU have elaborated and refined but not fundamentally challenged this formulation. There are many parallels here that strike a stu-dent of Indian politics: the “chief arena” of politics is the lower tier; the higher-level election is driven by political cleavages at the lower-level; and, independent electoral cycles of each of the “first order” elections play a crucial role in determining the “second order” verdict. It is tempting therefore to apply this model to the Indian reality and argue that the Lok Sabha election in India is like the European Parliament election and the national elections in Europe are like the state assembly elections in India. Political commentators and actors in India have intuitively grasped this point when they say that the Lok Sabha elections is 29 simultaneous but independent state level electoral races, or that the verdict in the Lok Sabha elections is nothing but an aggregation of state level verdicts.5 Such a simple application of the European model to the Indian context would however miss the real point: this comparison al-lows us to understand the specificity of the Indian experience. Itis not just that the logic of multiple but interlocked elections applies in India at the level of national elections. More impor-tantly, the interlocking follows a different logic here. The funda-mental difference between the European and the Indian cases is that the “second order” elections here are not devoid of real and substantive political significance. In India, a parliamentary election involves choosing a political executive that wields real political power, much wider in scope than the state level political executive. Thus, it is not that voters find the national election devoid of any substance, but that the substance is defined through the filters of state level issues, incumbents and strategies. It is not that the national government plays or is seen to be playing a secondary role. The institutional arrangement put in place by the Constitution loads the division of power in favour of the centralgovernment. Arguably, the central government has gainedinpower and economic clout precisely in the period when state politics has gained greater autonomy. Yet national politicsisnottheprincipal arena of political choices; political preferences and loyalties at the national level derive from primary loyalties in state politics. In that sense, unlike Europe, the logic of multiple but interlocked elections in India creates a disjunction between the level of effective political choice and the level of effective governance.In contemporary India, the principal arena of political contes-tation is at the state level. This is not simply a function of the state level being closer to the people than the national level. Elections to the third tier of democratic competition represented by pan-chayats and nagarpalikas, which are much closer to the people than state politics, do not shape political choices at the national or even state level.6 For complex historical reasons, state politics has come to be the point where micro aspiration and mobilisation from below meet the systemic requirements of macro politics. As such state level politics shapes and filters rather than pre-determines the national outcome. Political choices in a national election derive from the competitive format, electoral cycle, political agenda, participatory pattern, and social cleavage defined in state politics. In this sense, the political choices made at the state level are mostly “principal” and those made at the national level are increasingly becoming “derivative.” The “principal” nature of state level contestation has many aspects. One, the format of party competition and the menu of political choices in the Lok Sabha elections is decided by the state level party system, often with a subtle change. The choices in the national election are “simplified” further due to elimination of some of the smaller parties that are considered non-viable in national politics.7 This is in contrast to Europe where “fringe parties” tend to have greater salience in European Parliament elections. Two, the calendar of state politics provides the temporal context in which the Lok Sabha elections is set. Though there are two independent and usually non-concurrent electoral cycles – one for the state assembly and the other for the Lok Sabha – the politically relevant calendar is set to the cycle of state assembly elections and the tenure of state governments.

SPECIAL ARTICLEEconomic & Political Weekly EPW february 7, 200957Three, the leveland nature of political participation in the Lok Sabha election follows the pattern set in state politics, though the turnout in a national election is often a step lower than in a state election. Assembly elections make higher mobilisation possible as more actors with more direct stakes are involved. Four, politically relevant social cleavages are also defined in the arena of state politics. As in the case of the party system, the larger scale of the national election often results in a simplifica-tion or grand coalition making. But the building blocs of these coalitions are drawn from state politics. Five, the cultural-ideo-logical context of each state sets the agenda for the national election either by way of deciding the salience of various issues in the state or by determining how various national issues will be received in a state.8 Finally, while state politics does not exhaust the issues in national elections, the performance of state governments is one of the most salient questions driving voters in a national election. More often than not the Indian voter uses the Lok Sabha elections to pass a verdict on the state government. Unlike Europe, we do not witness an “expressive” vote: using the higher-level election strategically to send a warning to the ruling party at the more proximate level which the voters otherwise continue to remain loyal to. All these aspects make state level elections an arena of primary or principal choice and the choices made during a parlia-mentary election assume a derivative character.Political choices in a national election are derivative, but are neither mere duplication nor simply an aggregation of the choices in the state political arena. State politics provides the most sali-ent but not the only political context for voters. The performance of a national government and other national issues do matter to the voter, though these issues are filtered through the prism of state politics. Though the format of political competition comes from the state arena, the pattern of political preference shows subtle differences in the national elections.9 Sometimes there is a transfer of national issues to state politics,10 though this trend is weaker than the influence in the opposite direction. In some cases, when the ruling party in a state is not an important player in national politics and the important players at the national level are not relevant in state politics, the two levels of voting can be detached from one another. The emergence of the state as the principal arena of political contestation is linked to the inauguration of the post-Congress polity or what has been described as the third electoral system. In the first two decades after the inauguration of electoral democracy there was not much of differentiation between national and state level elections: the Congress used to win both with about the same margin. The beginning of the second electoral system and the rise of plebiscitary politics in the 1970s and 1980s witnessed a differentiation between national and state level pol-itics. Far from being derivative, national politics was the princi-pal arena that overshadowed state politics. State elections were reduced, with a generous misuse of Article 356, to an echo of the national verdict. It was only in the 1990s that the elections to state assemblies assumed autonomy and also the capacity to shape the outcome of the national elections. The electoral cycle at the state level was also effectively detached from that of national politics and state-specific political calendars were established in most parts of the country. A political calendar involves a predictable cycle of elections, a routine of political activity around it and some pat-ternofregime alternation. Earlier there was an official electoral cal-endar but it did not yield a meaningful political calendar. The timingofstate elections was tied to national political events, leading to large-scale dismissals of state governments and big rounds of as-sembly elections in 1972, 1977-78, 1980, 1984 and 1990. The sur-vival of state governments depended on the vagaries of national politics.After 1990 we have seen the emergence of state level electoral cycles independent of the disruptions at the centre, as during 1996-99. In some states a political calendar came into being before the 1990s: it began in Kerala in 1980, in Andhra Pradesh and Karnataka in 1983 and in Assam in 1985. But most other states, including the Hindi heartland, experienced a political calendar only with the rise of the third electoral system: Haryana in 1987, Uttar Pradesh and Tamil Nadu in 1989, Himachal Pradesh, Rajasthan, Delhi, Madhya Pradesh, Bihar, Gujarat and Orissa in 1990, Maharashtra in 1994-95 and Punjab in 1997. Some states like West Bengal, Tripura, Arunachal Pradesh and Sikkim still do not have a definite political calendar because one-party dominance does not allow a pattern of political alternation. In some states like Jammu and Kashmir, Manipur, Nagaland and Meghalaya, political competition is still too fluid to yield any political routines. The pattern of political calendars in contemporary India has had far-reaching consequences that are often not noticed. In most states the political calendar has acquired its own rhythm and logic and thus become an invisible part of the political “environ-ment”. This has set in motion a state-specific political routine often involving a set pattern of regime alternation, giving rise to expectations about when it is whose “turn” to come to power. Since the calendars of different states do not coincide, we now have a pattern of one or two major rounds of assembly elections each year. The net result of this complex and somewhat random criss-crossing of state level political calendars is that none of the major parties or national level coalitions is in a position to control most of the states at any given time. Coupled with the logic of the national verdict being a derivative verdict, it means that no party or coalition is in a position to win most of the states in a national election and thus secure a comfortable majority. Nor can a na-tional alliance be swept aside across the country.These developments gave a unique federal character to political competition, buttressed claims of autonomy of the states, strengthened regional pride, and thus deepened the process of national integration. But in the process, the ability of the national electorate to assess the performance and policies of the national government was eroded. The emergence of the state as an arena of political competition facilitated the democratic upsurge of the early 1990s; but as that decade came to a close, this upsurge was beginning to be contained within the systemic framework oftheduality of principal and derivative choices. In a sense, the Lok Sabha elections of 2004 marked the end of the transfor-mative potential of the third electoral system, yet its systemic attributes are very much in place and continue to shape political

SPECIAL ARTICLEfebruary 7, 2009 EPW Economic & Political Weekly58contestationat the state and national levels. The model of principal contestation at the state level and derivative choices at the national level is, thus, an outcome of third electoral system, without its transformative potential.2 PrincipalStateLevelContests,2004-08Let us turn from this general statement of the nature of linkage between state and national level elections to a consideration of the pattern seen since the last national election. We shall first propose a typology of principal state level contests and then examine the elections held in the last decade. Principal contests may be analysed in terms of two axes. First, what was the out-come in terms of the fate of the incumbent government: did it retain or lose power? Second, what was the direction and degree of change in popular preferences: did the election witness strong anti-incumbency, mild anti-incumbency or consolidation of the incumbent government? These are clearly not the only two salient dimensions. Any such contest takes place in the context of the party system at the state level, the alliances shaped at the time of the election, and the nature of political choices available to the voters. But the two axes explain the contest while the other factors constitute the context. These two axes yield a broad typology of principal state level contests. Each of the state elections can be categorised as continuity, change or reconfiguration. 2.1 ContinuityNotwithstanding the popular image of rampant anti-incumbency, we have seen quite a few state governments win a mandate again in the last five years. The retention of power by incumbent par-ties in Madhya Pradesh, Chhattisgarh and Delhi in the latest round added to a long list that includes Maharashtra, Assam, Orissa, West Bengal and Gujarat. Of course, there were differ-ences in the victories of these state governments; some won with handsome margins and consolidated their base in the state, but others won only a qualified extension with a somewhat reduced majority. There are also instances like Maharashtra in 2004 or Delhi more recently where the ruling party or coalition actually lost votes heavily without the loss translating into a gain for the main opposition. We might call this reluctant extension. As we mention below, the trend is increasingly in the direction of continuity. However,this does not necessarily indicate the emergence of an era of better gov-ernance; nor can there be a uniform explanation for the extensionsob-tained by different state governments: apart from the varying formats of party competition in the states, an ex-tension is also a function of a winning social coalition (as in Maharashtra), over-optimal coalition of two or more parties (Orissa and West Bengal), popularity of the state level leadership (Gujarat or Delhi) or simply the failure of the opposition to put its act together (Assam, Chhattisgarh, Madhya Pradesh). While the increased possibility of continuity gives political incentives to ruling parties, its long extension is associated with a narrowing of political choices for thevoters.2.2 VoteforChangeIn a competitive parliamentary system, one would expect this to be the most common choice through which popular disaffection is registered and/or alternation in the incumbent is effected. In fact, many states have now entered the alternation or rotation mode irrespective of parties, leadership and/or performance – Kerala being the most prominent example of this type. Tamil Nadu and Rajasthan are two other instances. Ordinarily, in such a routine change, the difference in vote percentage of the two main contestants is slender. But in cases like Andhra Pradesh in 2004, the swing of popular mood can be quite strong, almost as ifthe people are saying that a reconfiguration of the political competition in the state is necessary. Political commonsense associatesthis with non-performing governments and this is trueinamajority of cases. But the cases of Kerala, Tamil Nadu and Rajasthan show that this could well happen to reasonably well-functioning governments. In these cases it is a function of two fairly balanced political alternatives. Such alternation irre-spectiveof performance and/or policy indicates the narrowness of the difference between the two contestants. 2.3 ReconfigurationThis is not a simple case of parliamentary alternation. In this type the social bases of political forces change as happened in Uttar Pradesh in 2007. Or this could involve major changes in the party system, as in Karnataka recently. Another variant of reconfigura-tion would be a systemic shift that involves a consolidation of the bigger parties at the cost of the smaller parties. This opens the possibility of the defeat of the ruling party despite gaining votes. This happened in Punjab and Uttarakhand in 2007. It should be noted that this period did not see any classic instance of recon-figuration (as, say, in Andhra Pradesh in 1983, Haryana in 1987, Uttar Pradesh in 1993 or many states in north India in the 1990s) that characterised the third electoral system. The cases listed here in this category have at best an element of reconfiguration: sudden expansion of the social base of the Bahujan Samaj Party (BSP) in Uttar Pradesh, the entry of the Telangana Rashtra Samithi (TRS) in Andhra Pradesh, state-wise expansion of the BJP in Karnataka, or the changing regional and religious basis of voting in Jammu and Kashmir. That is why some of these Table 1: Typology of State Assembly Elections(2004-09)Consequences for Direction and Degree of Change in Popular Preferences Incumbent Government Swing for Ruling Party Moderate Swing against Ruling Party Big Swing against Ruling PartyContinuity for ruling Consolidation: WB2006, Qualified extension: MP 2008, Reluctant extension: party/coalition Nagaland 2008, Sikkim 2004 , Gujarat 2007, Orissa 2004 Assam2006,Arunachal2004, Tripura 2008, Pondicherry, Maharashtra 2004, Chhattisgarh 2008 Delhi 2008 Power shifts to Systemic shift: Punjab 2007 Vote for change: Rajasthan 2008, Vote of no confidence: the opposition Uttarakhand 2007 Goa 2007, Haryana 2005, Bihar 2005, Kerala 2006 HP 2008, Mizoram 2008, TN 2006, Jharkhand 2005New party or coalition Reconfiguration: AP 2004, comes to power UP 2007, Karnataka 2008 , J&K2008 Big swing is defined here as loss of 5 or more percentage points votes for the ruling party.

SPECIAL ARTICLEEconomic & Political Weekly EPW february 7, 200959instances do not display the features associated with reconfigu-ration: changing social cleavages, building of new social coali-tions and opening up of the political system.What do we find if we graft the verdicts of all the principal state level contests between 2004 and 2009 onto this typology? Table 1 (p 58) shows the broad patterns. The most striking thing about the period 2004-09 is the near absence of two categories that used to be very frequent in the early phase of the third elec-toralsystem: a vote of no-confidence and reconfiguration. But thisperiod saw very few truly “wave” elections. In contrast to the early phase of the third electoral system, this period saw a shift towards qualified extension or vote for change. Table 2 shows the rather dramatic decline in governmental anti-incumbency at the state level in the course of the last two decades. The early phase of the third electoral system was associated with a very high rate of attrition for the incumbent governments: 77% of the governments lost elections between 1989 and 1999. The rate has fallen sharply since then: between 1999 and 2003 it dropped to 62%, and in the last five years, it has come down to 46%. (Table 3 gives details of the state assembly elections in 1999-2003 and 2004-08.) These are signs of the routinisation of the system.Since the establishment of the third electoral system, the state level electoral cycles have taken various forms. The first and the most common type is a routine of regular oscillation from one party to another from one election to the next. This oscillation could take the form of the ruling party being thrown out in each election and being replaced by the dominant opposition party. What Ashis Nandy once called the “iron law of Indian democracy” was formulated in the context of national politics in the 1970s and 1980s but this could well describe the pattern at the state level. This involves a short honeymoon for the recently elected government, followed by disappointment and anger leading to “anti-incumbency” and the eventual defeat of the incumbent in the next election. Kerala has followed this pattern since 1980. Given the very high rate of attrition of incumbents in the 1990s, one can find many instances of the operation of this “iron law”; no wonder this has become the stereotype. But it is surprising to see how few states have actually followed this pattern: Kerala since 1980, Rajasthan since 1993, Punjab since 1992 and Himachal Pradesh since 1985. More common is the variant that may be called “irregular oscillation” where a change of guard is punctu-ated by “skips” as the ruling party sometimes manages to buck the trend; this category comprises Madhya Pradesh, Andhra Pradesh, Assam, Maharashtra and Delhi. There is another smaller variant of this pattern that can be described as “rotation” where power oscillates not between two major parties but rotates among more than two parties: Uttar Pradesh, Karnataka, and Haryana in recent times represent this category. And then there are states that have not entered this cycle due to continued one-party domi-nance (the Left Front in West Bengal and Tripura, the Congress in Arunachal Pradesh, the Sikkim Democratic Front in Sikkim, and theBJP in Gujarat). Or where the party system is too fluid to yield any pattern (Manipur, Meghalaya, Jammu and Kashmir).3 Derivative National Choices, 2004-09In order to appreciate the importance of state level verdicts as principal contests producing second order or derivative choices in the national contest, we now examine the various effects of the state-specific rhythm of an electoral cycle on the national election in that state. The national verdict could either lead to a reiteration of the state level outcome or its reversal. This is predicated on the point of time at which the national election takes place in relation to the state election. We can think of three stages in the state political calendar when national level choices may be made: early incumbency, mid-term or end-term evaluation. The derivative choices in the national election can thus be categorised as follows. 3.1 EarlyAssessment When a national election follows a state election in quick succes-sion, the state verdict tends to be reconfirmed in the national election from that state. The voters are ready to trust the parties Table 2: Incumbent Party's Performance in Assembly Elections (1989-2008) Total State Information Incumbent Won Incumbent Defeated ElectionsAvailableon 1989-98 63 57 1323% 4477%1999-2003 30 29 1138%1862%2004-08 30281554%1346%Those cases where it was difficult to determine who was the incumbent or who really won the election have been excluded as “No Information”.Source: CSDS Data Unit.Table 3: Trends in ‘Anti-incumbency’: Re-election of and Change in Votes for the Ruling Party in State Assembly Elections (1999-2008)States Did the Ruling Party Lose the Election? Change in Vote Share of Ruling Party 1999-2003 2004-08 1999-2003 2004-08 Andhra Pradesh No (TDP) Yes (TDP) -0.27 -6.28Arunachal Pradesh No (INC) No (INC) 1.28 -7.37Assam Yes (AGP) No (INC) -10.06 -8.67Bihar No (RJD) Yes (RJD in 2005 Oct) 0.72 -9.52Goa Yes (INC in 2002) Yes (BJP) -0.16 -5.25Gujarat No (BJP) No (BJP) 5.04 -0.73Haryana Yes (HVP) Yes (INLD) -17.11 -2.84Himachal Pradesh Yes (BJP) Yes (INC) -3.64 -2.1Jammu and Kashmir Yes (NC) Yes (INC+PDP) -6.54 -6.49Karnataka Yes (INC) Yes JD(S) -5.57 -1.65Kerala Yes (LDF) Yes (UDF) -2.4 -6.4Madhya Pradesh Yes (INC) No (BJP) -8.96 -4.86Maharashtra Yes (Shiv Sena in) No (INC+NC) 2.68 -9.99Manipur UnclearUnclear Meghalaya No(INC) Unclear -5.07..Mizoram No (MNF) Yes (MNF) 24.99 -1.04Nagaland Yes (INC) No (NPF) -14.88 3.86Orissa Yes (INC) No (BJD) -5.3 -2.04Punjab Yes (SAD+BJP) Yes (INC) -9.22 4.74Rajasthan Yes (INC) Yes (BJP) -9.3 -4.93Sikkim No (SDF) No (SDF) 10.32 18.77Tamil Nadu Yes (DMK) Yes (AIADMK) -11.15 1.2Tripura No (CPIM) No (CPIM) 1.33 1.19Uttar Pradesh Yes (BJP) Yes (SP) -12.04 0.06West Bengal No (CPIM) No (CPIM) -1.33 0.54Delhi No (INC) No (INC) 0.37 -7.82Pondicherry No (INC) No (INC) -2.56 7.11Jharkhand Yes(RJD+JMM) Yes (BJP) x -1.56Chhattisgarh Yes (INC) No (BJP) x 1.08Uttarakhand Yes (BJP) Yes (INC) x 2.68All-India Yes 18/29, Yes 15/28, -3.03 -1.73 No 10/29 No 13/28x state not in existence in 1999.

SPECIAL ARTICLEfebruary 7, 2009 EPW Economic & Political Weekly60they have elected at the state level and therefore use the occasion of the Lok Sabha elections to confirm the “probation” of the state government. This confirmation often comes with an “increment,” as the party that won the state assembly elections tends to win more Lok Sabha seats than its level of support in the assembly election would indicate. This is what happened in Rajasthan in 2004 as theBJP improved on its narrow victory in the assembly elections in 2003. In Madhya Pradesh and Delhi, however, the winner lost a little in terms of seats in the Lok Sabha elections. This possibility would encourage us to believe that parties elected to state governments close to within one year from a Lok Sabha elections stand a fair chance of performing well or in fact improv-ing their performance from those states. 3.2 Mid-termReviewOn the contrary, if a national election is held at a time that hap-pens to be the middle of the term of an elected government in a state or later, voters are more likely to adopt a critical approach towards the parties in power at the state level. Rather than un-questioning confirmation, this often involves a “suspension notice”. The Lok Sabha elections of 2004 in Kerala, Punjab and Haryana were good examples of this: the ruling party fared very badly and went on to lose the next assembly election. The “derivative” choice thus is not merely a protest vote; apart from restricting the fortunes of contesting parties from that state, this vote also an-ticipates the verdict in state elections to be held later. But the mid-term review need not result in a negative assessment. State governments that have or are likely to get an “extension” are likely to win a provisional approval in the national election as well. Assam, Maharashtra and West Bengal in 2004 illustrate this pattern. However, Uttar Pradesh turned out to be an extension for the Samajwadi Party – the party strength in the Lok Sabha reflected the outcome of the previous assembly elections – but the outcome did not anticipate the assembly elections verdict in 2007.3.3 Final VerdictIn a similar fashion, if a national election takes place when a state election is round the corner or is taking place simultaneously, the two choices are likely to merge with each other. Indian voters rarely split their votes so much as to give a different verdict at the state and the national elections if they are held simultaneously or nearly so. The verdict could be a final approval or a dismissal and is replicated at the national level in about the same proportion. Both Andhra Pradesh and Orissa had simultaneous elections to the Lok Sabha and state assembly in 2004. In both places, a simple re-aggregation of the state assembly outcome fitted the actual outcome of the Lok Sabha elections. We can finally turn to the linkage between state elections held between 2004 and 2008 and the coming Lok Sabha elections of 2009. If Lok Sabha contests were simply a repetition of previous assembly verdicts, all we need to do is re-aggregate the assembly verdicts into Lok Sabha seats. This is what Table 4 (p 61) does: the votes for all the major parties in all the assembly segments of a Lok Sabha constituency have been added to impute a winner in a hypothetical election. This computation has been done for pre-delimitation constituencies in those states that went to polls before the new delimitation. Clearly this is not and is not meant to be a good guide to the Lok Sabha elections. Table 4 also shows the distance between the mechanical re-aggregation and the ac-tual result of the 2004 Lok Sabha elections, indicating that the national verdict is derivative but not repetitive. The linkage be-tween the principal and derivative choices is quite complex: while there is a transfer of political choice from the state to the national level, there is no way one can imagine state level choices being mechanically repeated at the national level. Against this backdrop, let us take a quick inventory of the elec-toral cycle through which the various states are now passing. Per-haps the easiest thing is to list the states that are likely to witness a confirmatory choice: Jammu and Kashmir, Rajasthan, Delhi, Mizoram, Nagaland, Pondicherry, Karnataka, Madhya Pradesh, Chhattisgarh and perhaps Himachal Pradesh, Tripura and Meghalaya. In this group comprising 116 Lok Sabha seats, the BJP has a slight advantage because 72 of these seats are in states where the party will press its probationer’s advantage. The real battleground will be the group of states facing a mid-term review. This is the largest category, comprising 281 seats, a little over a majority in the Lok Sabha. Here 84 seats are in states that are under National Democratic Alliance (NDA) rule (Uttara-khand, Bihar, Gujarat, Punjab), while United Progressive Alliance (UPA) states account for just 55 seats (Tamil Nadu, Assam, Goa), leaving as many as 142 seats in states (Uttar Pradesh, Kerala, West Bengal) where the ruling parties at the moment are not part of either of these two coalitions. As argued above, electoral pos-sibilities are wide open in this group of states. Earlier one could have assumed a tilt against the ruling party, but this assumption is difficult to sustain now in the light of the decline in incum-bency disadvantage. Awaiting final approval or disapproval of the electorate are the states of Andhra Pradesh, Haryana, Sikkim, Orissa, Jharkhand, Maharashtra and Arunachal Pradesh, which account for 137 seats in the Lok Sabha. Not all the states in this category go to the polls with the Lok Sabha; in many of these states elections take place within a year of the Lok Sabha elections. In this group, the UPA has the biggest stakes; Orissa is the only NDA-ruled state here. Since these are not routine-oscillation states, it is not easy to project the outcome simply on the basis of past performance and pattern. At least three of these states involve considerable uncer-tainty: while it will be a test election for a significant and new political formation in Andhra Pradesh (Chiranjivi’s Praja Rajyam) and Haryana (Bhajan Lal’s Haryana Janhit Congress) that could lead to a reconfiguration of the party system, in Jharkhand the political mess in the recent past could invite an unpredictable response. Both in Maharashtra and Orissa, the ruling party/ coalition has a long incumbency to defend, though it cannot be presumed that this alone would be a decisive liability. The purpose of this listing is not to project the final outcome, but only to understand the nature of contestation in terms of the interlocking of the state and the national elections. The outcome in each state will depend, apart from the location in the state’s electoral cycle, on the intensity of popularity/unpopularity of the ruling parties, the shape of pre-election alliances and the other political contingencies of the next few weeks. If performance of

SPECIAL ARTICLEEconomic & Political Weekly EPW february 7, 200961the ruling parties or coalitions becomes an important factor in shaping the outcome of the election, then the coming Lok Sabha election are going to be more of a test for the UPA than for the NDA. The performance of the union government will also be an issue, though this issue will be viewed in the prism of state poli-tics. The UPA government led by Manmohan Singh started with-out a clear popular mandate. The leadership of the government wasalsoseen as lacking in legitimacy. Gradually, the UPA ac-quired acceptance and even popularity. The State of Nation Sur-veys conducted by the CSDS in 2006 and 2007 show that the popularity of the UPA government declined considerably. As the UPA government prepares to face the electorate, it appears to be facing dissatisfaction. The real question is whether this rise and then fall in the popularity of theUPA leaves it at a point higher than its starting point in 2004. The complex pattern of principal outcomes and timing in the political calendar shows why neither of the two major national alliances can sweep the polls nor be swept aside in the coming election. In all probability, we are going to witness one more election in which no single party or pre-poll alliance is likely to get a clear majority in the Lok Sabha. That leaves open a wide range of possibilities, perhaps wider than was the case in 2004, of government formation in the post-election scenario. In such a situation, where every seat matters, the smallest of shifts can make a big difference. Shifts in popular choice in the big and uncertain states like Tamil Nadu, Andhra Pradesh, West Bengal, Bihar and Uttar Pradesh will be decisive in determining the final outcome.Making a precise projection of the outcome is neither possible at this stage nor does it fall within the purview of this essay. The point here is about understanding the structure of contesta-tion that will shape the final outcome. The complex temporality forced by electoral cycles specific to different states shaping the principal contests at the state level, the continuous public Table 4: Lok Sabha Seats Won or Led by Major Parties by States(1999-2008)State Actual Outcome Lok Sabha Imputed Outcome Assembly Elections Actual Outcome Lok Sabha Imputed Outcome Assembly Elections Elections 1999 (Seats Won) 1999-2003 (Lok Sabha Seats Led) Elections 2004 (Seats Won) 2004-08 (Lok Sabha Seats Led) Total INC BJP Others Assembly INC+ BJP+ Others Congress+ BJP+ Others Assembly Congress+ BJP+ Others SeatsAlliesAllies ElectionAlliesAllies AlliesAllies ElectionAlliesAllies YearYearAndhra Pradesh 42 5 36 MIM-1 1999 5 37 -34 5 LF 2 MIM-1 2004 30 11 MIM-1Arunachal Pradesh 2 2 --1999 2 ---2 -2004 2 --Assam 14 10 2 IPF-1 IND-1 2001 13 -IND-1 9 3 AGP-2 2006 9 -AGP+-3 AUDF-2Bihar 40 2 30 1 RJD-6 IND-1 2000 1 17 RJD-22 29 11 -Oct-2005 9 30 LJNSP+-1Chhattisgarh 113 8 -2003 4 7 -1 1020084 7 Delhi 7 -7 -2003 7 --6 1 20087 Goa 2 -2 -20021 1 -1 1 -20071 1 -Gujarat 26620-2002224-1214-2007917-Haryana 10 -10 BJP-5, -2000 4 6 -9 1 -2005 9 -INLD(R)-1 INLD-5(BJP,INLD) Himachal Pradesh 4 -3 HVC-1 2003 3 1 -3 1 -2007 1 3 -Jammu and Kashmir 6 -2 JKNC-4 2002 2 -JKNC-2 4 -JKNC-2 2008 1 1 JKNC-2 PDP-1 PDP-1 IND-1 Oth 1Jharkhand 14 2 11 RJD-1 2000 1 9 RJD-2 JMM-2 13 1 -2005 -8 RJD-1 JMM-2 Oth3Karnataka 28 1810-199926 2 -8 18JD(S)-2 2004149 JD(S)-5Kerala 20 11 -LF 9 2001 16 -LF 4 1 1 LF 18 2006 4 -LF 16Madhya Pradesh 29 8 21 -2003 -29 -4 25 2008 7 22 Maharashtra 48 11 28 NCP-6 PWP-1 1999 11 27 NCP-10 23 25 - 2004 33 15 - Other2 Manipur 2 --NCP-1 20022 --1 -IND-1 20072 -- MSCP-1 Meghalaya 21-NCP-120031-NCP-111-20081-NCP-1Mizoram 1 --IND-1 2003 --MNF-1MNF-120081 Nagaland 11--20031---NPF-1-20081--Orissa 212 18 JMM-1 20001 20 -2 18JMM-1 2004 3 18 -Rajasthan 259 16-2003 7 18-4 21200813111Punjab 13 8 3 LF 1 SAD(M)-1 2002 8 5 -2 11 -2007 2 11 -Sikkim 1 -1 (SDF) -1999 -1 (SDF) --1 -2004 --SDF-1Tamil Nadu 39 12 26 1 2001 30 9 35 - 4 2006 27 - AIADMK+-12Tripura 2 --2 2003 ---LF-220082Uttarakhand 5 1 4 -20023 1 BSP-1 1 3 SP-120071 3 BSP-1Uttar Pradesh 80 9 25 BSP-14 SP-26 2002 3 10 BSP-23 SP-42 9 11 BSP-19 2007 2 7 BSP-52 SP-18 RLD-2 Oth-4 RLD-1 Oth-1 SP+38 Oth 3 Oth 1West Bengal 42 3 10 29 2001 --LF 39 TRMC-3 6 1 LF 35 2006 1 1 LF 40Computations for “imputed outcome” were made by aggregating the votes secured by major parties in all the assembly segments falling within a Lok Sabha constituency (as per the prevailing delimitation at that time). The party that secured highest votes was treated as having “led” in that parliamentary constituency and was “awarded” that seat.Source: CSDS Data Unit.

SPECIAL ARTICLEfebruary 7, 2009 EPW Economic & Political Weekly62scrutiny faced by the central government along with the deriva-tivecharacter of the parliamentary choices, and the peculiar trajec-tory of public approval of the UPA government provide us with the backdrop against which the popular choices in the 15th Lok Sabha elections will unfold. 4 ConclusionsWhat are the political consequences of this specifically Indian model of multiple, differentiated yet interlocked elections? This largely unexpected, unintended and under-theorised pheno-menon has a wide range of consequences that we may have only begun to understand. On the one hand, a complex political calendar and transfer of political choices from the state to the national level makes for greater accountability at both levels of government. In the last decade, we have got used to one major round of state assembly elections every year; putting the central government under continuous assessment. At the same time, the state governments cannot solely rely on the national gov-ernment for ensuring their victory or blame it for their defeats. In this sense, the state governments are more vulnerable and accountable today than in the days of nation-wide electoral waves. This has also led to much greater political inclusiveness than has ever been the case since the beginning of democratic politics in India. There is almost no major political party at a given point that is not part of a government either at the state or the national level, or has not been part of these governments in the recent past. As noted above, this has also meant greater regional integration as parties from the hitherto peripheral states have come to play a crucial, sometimes pivotal, role in the formation of government at the centre. The continuing role of Dravida political parties in the union governments since 1996 is a case in point.On the other hand, the differential electoral cycles at state and national levels and the increasing derivative character of national level choices have had an unnoticed consequence: a disjunction between the level of accountability and public scrutiny at the state level and the real centre of power at the national government. This disjunction gives the national government a unique insularity or autonomy – its fortunes are less dependent on what it does or does not do, and more on how the state governments do and how that fits into the political calendar. The politically inclusive character of the system has a flip side: the third electoral system has in effect produced the political establishment that leaves little space for a real opposition. Elsewhere we have already elaborated on the party competition based on “convergence”.11 We suggest that the political mechanism that produces the convergence can be better understood if we take into account this linkage between the principal and derivative arenas of political choices. If the 1990s drew our attention to the possibilities of social transformation through political reconfiguration within competi-tive politics in a multiparty framework, the present decade alerts us to the tendency within that same framework to tame the democratic upsurge and contain the transformative potential of competitive politics. This presents us with multiple paradoxes: as the frequency of accountability goes up, the scope of account-ability gets narrower than before; as the possibility of a stable parliamentary majority recedes, it leads to greater stability of policies that can be kept out of the democratic framework; the higher the suspense over who will eventually form the govern-ment, the lesser it matters. Notes 1 Based on its post-poll surveys, the Lokniti Team has reported many of its findings of state level elections: “How Assam Voted”,The Hindu, 18 May 2006; “How Bihar Voted: Assembly Elections 2005”, The Hindu, 10 March 2005; “How Haryana Voted: Assembly Election 2005”,The Hindu, 6 March 2005; also, see: Abhay Datar, “How Maharashtra Voted: Of Regional Variations and Shifting Strongholds”, The Hindu, 24 October 2004; Rajeshwari Deshpande and Nitin Birmal, “Social Chemistry and Gender Did the Trick”, The Hindu, 24 October 2004; Sanjay Kumar, “Maya ‘Class’ Act Has Upper Castes Split”, Indian Express, 2 May 2007; Sandeep Shastri, Sanjay Kumar, “Electoral Flavors of Different Regions in State”, Deccan Herald, 23 May 2008; Yogendra Yadav, “A Story of Political and Ethnic Fragmentation”,The Hindu (Delhi), 18 May 2006; Yogendra Yadav, “How Kerala Voted”, The Hindu (Delhi), 17 May 2006; Yogendra Yadav, “More to It Than Meets the Eye”, 17 May 2006; Yogendra Yadav and Sanjay Kumar, “Not Alliance Arithmetic Alone: How Jharkhand Voted”, The Hindu, 7 March 2005; Yogendra Yadav and Sanjay Kumar, “Poor Man’s Rainbow Over UP”, Indian Express, 17 May 2007; Yogendra Yadav and Sanjay Kumar, “Present Imperfect”,Indian Express, 22 May 2007; Yogendra Yadav and Sanjay Kumar, “Shak-en to the Core”,The Indian Express, 18 May 2007; Yogendra Yadav and Sanjay Kumar, “Traces of a Mandate”,Indian Express, 23 May 2007. 2 This argument was first made in Yogendra Yadav (1996), “Reconfiguration in Indian Politics: State Assembly Elections, 1993-95”, Economic & Political Weekly, Vol 31, Nos 2 and 3; 13-20 Janu-ary, 95-104. 3 Yogendra Yadav (1999), “India’s Third Electoral System”, Economic & Political Weekly, Vol 34, Nos 34-35, 2393-99. 4 Karlheinz Reif and Hermann Shmitt (1980), “Nine Second-Order National Elections: A Conceptual Framework for the Analysis of European Election Results”, European Journal of Political Research, 1980, 8 (1): 3-44. 5 For instance, see this interview with L K Advani: http://timesofindia.indiatimes.com/articleshow/ 706300.cms and http://www.indianexpress.com/oldstory.php?storyid=47902<http://www.express-india.com/news/fullstory.php?newsid=31894 6 It is common for the media to draw on the results of the latest panchayat or municipality elections in the state as “straws in the wind” to gauge the public mood in the next assembly elections. Yet this method has proved a very poor guide to elec-tion outcomes. Recently, the Congress was swept away in the Municipal Corporation of Delhi elec-tion before being voted back to power in the assembly elections in 2008.7 This would be true of smaller parties like the Apna Dal in Uttar Pradesh, the Gondawana Gantantra Parishad in Madhya Pradesh, the Maharashtra Ekikaran Samiti in Karnataka and the Jharkhand Party in Orissa that fall off the map as a result of the upscaling in the Lok Sabha elections. The United Democratic Front and the Left Democratic Front in Kerala, the Left Front in West Bengal, the third front in Maharashtra and the Dravida Munnetra Kazhagam and the All-India Anna DMK-led fronts in Tamil Nadu also tend to drop manysmallerpartnersinthe Lok Sabha elections.8 The popular reactions to global economic melt-down in Punjab, international terrorism in Gujarat and foreign policy towards Sri Lanka in Tamil Nadu are instances of how seemingly remote issues can lead to substantive and differ-ential receptions in different states. We discuss some of these state-specific patterns in Yogendra Yadav and Suhas Palshikar (2008), “Ten Theses on State Politics”,Seminar, November, No 591, 14-22.9Sometimes this difference is systematic. For example the Congress Party does better in Lok Sabha elections than in state assembly elections in Uttar Pradesh. The same is true of the BJP in Karnataka and Assam and the UDF in Kerala.10 The state assembly elections in Maharashtra in 2004 witnessed the ruling Congress-Nationalist Congress Party alliance benefit from the honey-moon period of the UPA government at the centre. In some of the smaller states, especially the hill states in the north-east, state elections are influ-enced by an appeal to vote the party that is in sync with the ruling party at the centre so as to get better central patronage.11 We argue elsewhere that an expansion in the format of political competition has been accom-panied by a shrinkage in the nature of policy choices available to a voter: Yogendra Yadav and Suhas Palshikar (2003), “From Hegemony to Con-vergence: Party System and Electoral Competi-tion in the Indian States, 1952-2002”, Journal of Indian School of Political Economy, Vol 15, Nos 1 and 2, 5-44.