UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Party politics in a non-western democracy : a test of competing theories of party system change, government… Nikolenyi, Csaba 2000

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata


831-ubc_2000-486842.pdf [ 15.18MB ]
JSON: 831-1.0089643.json
JSON-LD: 831-1.0089643-ld.json
RDF/XML (Pretty): 831-1.0089643-rdf.xml
RDF/JSON: 831-1.0089643-rdf.json
Turtle: 831-1.0089643-turtle.txt
N-Triples: 831-1.0089643-rdf-ntriples.txt
Original Record: 831-1.0089643-source.json
Full Text

Full Text

PARTY POLITICS IN A NON-WESTERN DEMOCRACY: A TEST OF COMPETING THEORIES OF PARTY SYSTEM CHANGE, GOVERNMENT FORMATION AND GOVERNMENT STABILITY IN INDIA by . CSABA NIKOLENYI B . A . , Memorial University of Newfoundland , 1993 M . A . , The University of British Columbia, 1994 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES ( Department of Political Science ) We accept this thesis as conforming to the required standard  THE UNIVERSITY OF BRITISH COLUMBIA  2000 © Csaba Nikolenyi, 2000  In  presenting  degree freely  at  this  the  available  copying  of  department publication  of  in  partial  fulfilment  University  of  British  Columbia,  for  this or  thesis  reference and  thesis by  this  for  his thesis  scholarly  or for  her  of  The University of British Vancouver, Canada  Date  DE-6  (2/88)  A c c ^ c A .  JO  Columbia  I  I further  purposes  gain  the  shall  requirements  agree  that  agree  may  representatives.  financial  permission.  Department  study.  of  be  It not  is  that  the  Library  permission  granted  by  understood be  for  allowed  an  advanced  shall for  the that  without  make  it  extensive  head  of  my  copying  or  my  written  II  ABSTRACT The dissertation will address the ongoing debate in Comparative Politics about the virtues and pathologies of rational choice theory by testing competing hypotheses and predictions to account for three aspects of party politics in India: the transformation of the Indian political party system from a predominant to an even multiparty system; the politics of government formation; and the politics of government stability. Overall, the dissertation will pursue two arguments. First, rational choice models and predictions can account for the empirical cases more consistently than hypotheses and predictions derived from other paradigms. Second, by using India as the case on which to test competing theories, it will be shown that non-Western political phenomena are not sui generis and they may be accounted for in terms of comparative theory the same way as Western phenomena have been.  iii  ABSTRACT  ii  LIST OF T A B L E S  vii  LIST OF FIGURES  ix  DEDICATION  xi  ACKNOWLEDGEMENT  xii  C H A P T E R 1: INTRODUCTION 1. 1 The Theoretical Problem: Rational Choice and Its Critics  1  1 . 2 The Empirical Problem: Party System Change, Government Formation, and Government Stability in India  7  1 . 3 Methodology  13  1 . 4 Comparative Studies on the Indian Party System and Coalition Politics  16  C H A P T E R 2: P A R T Y POLITICS IN INDEPENDENT INDIA 2 . 1 The First Three Lok Sabhas, 1952-67: The Dominance of the Congress Party  21  2 . 2 The Fourth Lok Sabha, 1967-71: Split in the Governing Party  23  2. 3 The Fifth Lok Sabha, 1971-77: Emergency Rule and Its Consequences  39  2 . 4 The Sixth Lok Sabha, 1977-80: Two Unstable Coalition Governments  47  IV  2.5 The Seventh and Eighth Lok Sabhas, 1980-89: Congress Dominance Restored  59  2.6  The Ninth Lok Sabha, 1989-91: The Janata Failure Repeated  69  2.7  The Tenth Lok Sabha, 1991-96: A Stable Minority Government  81  2.8  The Eleventh Lok Sabha, 1996-97: Three Unstable Minority Governments  CHAPTER 3: PARTY SYSTEM CHANGE 3.1  Conceptualizing the Dependent Variable  87 102 103  3 . 2 The Conventional Wisdom of the Indologist Literature  110  3 . 3 Testing Conventional Wisdom  116  3.4  Comparative Theories of Party System Change  123  3.5  A Test of Two Theories  128  3 . 6 Strategic Voting and Linkage  130  3.7  A Strategic Cooperation Explanation of Changes in the Indian Party System  3 . 8 Conclusion CHAPTER 4: GOVERNMENT FORMATION 4.1  Minimum Winning Coalition Theories  141 153 154 154  4 . 2 Party System Characteristics and Government Types  157  4 . 3 Macro-Political Regime Characteristics  163  4.4  Opposition Influence and Electoral Decisiveness  168  4.5  The Core  175  4.6  Dominant and Central Players  178  4 . 7 Constraints and Cabinet Formation  185  4 . 8 Constraints on Coalitional Bargaining in India  189  4 . 9 Government Formation in India  193  4 . 1 0 Conclusion  214  C H A P T E R 5: G O V E R N M E N T STABILITY  5.1  Time-Dependent Explanations  215  216  5 . 2 Time-Independent Explanations  218  5.2.1  The Balance Among Parties in the Legislature  218  5.2.2  Institutional Features of the Political System  226  5.2.3  Ideology  227  5.2.4  Key Players  229  5.2.5  Key Players and the Transformation of the Parliament  239  5 . 3 Conclusion  C H A P T E R 6: CONCLUSION: THE NESTED G A M E S OF P A R T Y POLITICS PN INDIA  6.1  A Brief Overview of the Findings  241  243  244  6 . 2 Unifying the Themes: The Nested Games Approach  249  6 . 3 The Nested Games of Party Politics in India  260  BIBLIOGRAPHY A P P E N D I X 1: LIST OF PARTIES MENTIONED fN THE T E X T  277  APPENDIX 2: SIMPLE GAME THEORY APPENDIX 3: THE SCHOFEELD ET AL. THEOREM APPENDIX 4: CLASSIFICATION OF LOK SABHAS APPENDIX 5: CONSTRAINTS AND THEIR IMPACTS  Vll  LIST OF T A B L E S  1.1  Party System Characteristics in Four Westminster Democracies, 1945-97  1.2  Parliaments and Governments in Four Westminster Democracies, 1945-97  2.1  10  The Position of the Congress Party and the Opposition in the 1969 Presidential Electoral College  2 .2  9  28  The Changing position of the Congress Party in the Fourth Lok Sabha, 1967-71  32  2.3  The Party-Wise Composition of the Rajya Sabha, 1966-70  36  2 .4  Concurrent and Non-Concurrent National and Subnational Electons in India, 1952-96  2.5  39  The Sixth Lok Sabha, 1977-80, Before and After the Split of the Janata Party  47  2 .6  The Party-Wise Composition of the Rajya Sabha as of March 1976  48  2 .7  Electoral Coordination between the BJP and the Janata Dal in 1989 and 1990  2 .8  73  Number and Percentage of Seats Won by Party by Muslim Population Proportion in the 1989 Lok Sabha Elections in Uttar Pradesh  2 .9  The Impact of Rajiv Gandhi's Assassination on the Electoral Performance of the Major Parties in the 1991  76  viii  Lok Sabha Elections  82  3.1  The Ideological Location of Indian Political Parties  142  3 .2  Electoral Coordination Against the Congress Party, 1962-96  143  3.3  Pilihibit Constituency in 1971 and 1977  145  3.4  Mahasamund Constituency in 1984 and 1989  145  3.5  Agra Constituency, 1977-91  147  3.6  Bombay North Constituency, 1977 and 1980  148  3.7  Rajapur Constituency, 1977-91  150  3.8  Begusarai Constituency, 1977-91  151  3 .9  The Recovery and Decline of the Congress Party  152  4.1  Characteristics of Indian Governments, 1952-97  156  4.2  Ideological Polarization in the Lok Sabha, 1989-96  159  4 .3  Polarization in the Indian Party System, 1952-96  160  4 .4  Parliamentary Standing Committees in India, 1993  172  4.5  The Electoral Responsiveness of Government Formation in India  174  5.1  Party System and Government Characteristics in India, 1952-96  220  Appendix 4 . ICehtrality in the Fourth Lok Sabha, 1969  287  Appendix 4 . 2Centrality in the Sixth Lok Sabha, 1979-80  288  Appendix 4 . 3Centrality in the Ninth Lok Sabha, 1989-91  290  Appendix 4 . 4Centrality in the Tenth Lok Sabha, 1991 -96  291  Appendix 4 . 5Centrality in the Eleventh Lok Sabha, 1996-7  292  LIST OF FIGURES  3.1  The Congress Party in the Indian Party System, 1952-96  3.2  Sartori' Typology of Political Party Systems  3 .3  The Evolution of the Indian Party System in Terms of Sartori's Model  3 .4  The Fragmentation and the Effective Number of Parties in the Indian Party System, 1952-96  3 .5  The Evolution of the Indian Party System in Terms of Ware' Classification  3 .6  The Regional Breakdown of the Congress Party' s Electoral Performance, 1952-96  3 .7  The Congress Party' s Electoral Performance in the Region, 1952-91  3.8  The Age Structure of the Indian Electorate and the Electoral ft  Performance of the Congress Party 3 .9  The Actual and Mobilized Vote of the Congress and the Opposition Parties, 1952-96  3.10  The Actual and Mobilized Vote of the Congress and the Opposition Parties in the Hindi Heartland, 1952-84  3.11  The Actual and Mobilized Vote of the Congress and the Opposition Parties in the Non-Hindi Region, 1952-84  3.12  Electoral Volatility in India, 1952-96  3 . 13 The Tie Curve  137  3 . 14 Possible Transformations of Center-Dominated Multiparty Systems 3.15  138  The Performance of the Congress Party and the Effectiveness of Non-Congress Coordination  143  4.1  Dodd's Hypotheses About Cabinet Formation  158  4 .2  Fragmentation and Polarization in the Lok Sabha, 1952-96  161  4 .3  Seat Volatility in India, 1952-96  162  4.4,  Luebbert's Typology of Political Systems  164  4 .5  Voter Turnout in National Elections, 1952-91  166  4 .6  Strom's Typology of Political Systems  168  4 .7  Van Roozendaal's Typology of Dominated and Centralized Parliamentary Games  182  4 .8  The Classification of India's Hung Parliaments  184  4 .9  Cabinet Viability in a Four-Party System  187  5.1  Dodd's and Grofman's Theories About the Relationship Between Cabinet Type and Cabinet Durability  224  6.1  A Two-Dimensional Simplex of Cooperation in a Three-Player Game  359  6.2  The Performance of the Congress Party in By-Elections  264  Appendix 3.1 The Core at A When W=l and q =2/3  282  Appendix 3.2 The Core at A When W=2 and q= 2/3  283  Appendix 3.3 The Core at B When W=l and q= 1/2  284  Appendix 3 . 4 Empty Core When W=2 and q = 1/2  284  xi  ANDIKAMNAK  XII  ACKNOWLEDGEMENTS This work is the product of years of questioning, excitement, frustration, and, above all, enthusiasm. That my questions were answered and new ones were constantly raised; that my excitement was incessantly shared and fueled; that my frustrations found sympathetic ears; and that my enthusiasm was encouraged is thanks to a large number of people I have had the good fortune to interact with both professionally and in my private life. While the responsibility for the views expressed in this work rests solely with me, I am grateful for the supportive environment provided by the Department of Political Science of the University of British Columbia that enabled and encouraged me to engage in research the result of which is this dissertation. I am especially thankful to the members of my dissertation committee, Dr. John R. Wood, Dr. Ken R. Carty and Dr. Richard Johnston, for their incessant support; Dr. Masaru Kohno for his kind assistance; and Dr. Steven Wolinetz and Dr. Mildred Schwartz for always being there for me. Ms. Nancy Mina, Ms. Petula Muller and Ms. Dory Urbano never failed to help in times of need. Last but far from least, I remain forever indebted to my wife, Andrea, and to my inlaws, Drs. George and Csilla Gondos, for having so much trust in me. The generous funding of the Social Science and Humanities Research Council of Canada, the Shastri Indo-Canadian Institute, and the University of British Columbia is thankfully acknowledged.  1  CHAPTER ONE: INTRODUCTION THE THEORETICAL PROBLEM: RATIONAL CHOICE AND ITS CRITICS The economic approach to politics, also known as rational choice, has gained tremendous popularity in the recent comparative politics literature. While in 1952 there was not a single article in the rational choice vein published in the leading journal of the profession, the American Political Science Review, by 1992 approximately 40% of the articles published in the journal were written in this paradigm. However, rational choice 1  has attracted just as much criticism as it has gained followers since the publication of its foundational theoretical works by Hotelling , Arrow , Downs , Buchanan and Tullock , 2  3  4  5  Black , and Olson . 6  7  The distinctive characteristic of rational choice theory is the assumption that there is an optimal correspondence between the ends that individual political actors choose to pursue and the means they choose to pursue those ends with. In contrast, the earlier 8  dominant paradigms in comparative politics have either ignored the importance of actors or assumed the absence of any optimal correspondence between means and ends. Examples of the former include systems analysis, functionalism, structuralism and modernization theory that explain social and political phenomena by reference to the  'Donald P. Green and Ian Shapiro, Pathologies of Rational Choice Theory: A Critique Applications in Political Science (New Haven: Yale University Press, 1994), p. 6. Harold Hotelling, "Stability and Competition," Economic Journal 39 (1929), pp. 41-57. Kenneth Arrow, Social Values and Individual Values (New York: John Wiley & Sons, 3  1957). Anthony Downs, An Economic Theory ofDemocracy (New York: Harper & Row, 1957). James Buchanan and Gordon Tullock, The Calculus of Consent (Ann Arbor: University of Michigan Press, 1962). Duncan Black, The Theory of Committees and Elections (Cambridge: Cambridge University Press, 1958). Mancur Olson, The Logic of Collective Action (Cambridge: Harvard University Press, 1965). 'George Tsebelis, Nested Games: Rational Choice in Comparative Politics (Berkel University of California Press, 1990), p. 18. 4  5  6  7  2  system, howsoever it is denned, as a whole. While individual rationality is not necessarily denied by these macroscopic theories, it is ignored as irrelevant an unimportant. Theories that are interested in the individual but deny the significance o f 9  rationality locate the source of individual actions either in impulsive motives, such as relative deprivation , or "theoretical constructs, which may be inaccessible to both the 10  observer and the actor" , such as national culture or inertia. 11  Reduced to the very basics, rational choice theory rests on the following assumptions: •  political actors pursue certain goals that they set for themselves;  •  the goals set by political actors reflect the actors' self-interest;  •  political actors are conscious o f the choices they are making;  •  the individual is the basic political actor;  •  the ordering o f the actor's preferences are both consistent and display some degree o f stability over time;  •  when presented with a range o f options to choose from, individual political actors choose the alternative that is expected to yield the highest utility in terms o f the actor's preferences;  •  actors have information about the alternatives available to them as well as the consequences o f their choices among these alternatives.  12  It must be emphasized that the preferences of a rational actor must be both noncontradictory and transitive. Rationality is incompatible with contradictory preferences, where contradiction means the conjunction of a proposition and its negation, because they reflect the actor's inability to reason. If actors held contradictory preferences then it Tsebelis, Nested Games, p. 19-20. T e d Robert Gurr, Why Men Rebel (Princeton: Princeton University Press, 1971). "Tsebelis, Nested Games, p. 22. 10  12  Kristen Renwick Monroe, " T h e Theory of Rational Action" in K. R. Monroe, ed., The  Economic Approach to Politics: A Critical Reassessment of the Theory ofRational Actio (New York: Harper Collins, 1991), p. 4.  3  would be impossible tofindan optimal correspondence between their means and ends because they could choose any, rather than the optimal, tool to attain their goal. Transitive preferences simply mean that given three options, a, b, and c, a rational actor will prefer a to c if he/she prefers a to b and b to c. If an actor did not hold transitive preferences then it would be impossible for him/her tofindthe optimal, or least costly, means to attain his/her goal. For instance, suppose that an actor preferred a to b, b to c but c to a. This actor would be willing to give up option a in favor of c and incur the costs involved, however, he/she would also be willing to further trade option c for b, and again b for a. By the end, the actor has incurred significant costs, yet ended up with the same option as he/she started out with, i.e. option a. In contrast, if the actor held transitive preferences, then he/she is rational and would not incur these costs in vain. Provided that he/she prefers a to b, b to c, and a to c, the actor will not give up a once attained because there is no other option preferred to it. Practitioners of rational choice theory stress the methodological advantages inherent in this paradigm. According to Ferejohn, the assumption of utility maximization forces the analyst to come up with systematic rather than ad hoc explanations. For 13  example, suppose that an actor has to make a choice from the following set of alternative options {a^ a , • \}- According to the principle of utility maximization, an explanation 2  has to show why the actor prefers a to a n  n+1  for any n. However, if the utility  maximization principle is not satisfied then the actor may prefer a to a n  n+1  but also a to n  a^ In this case, the actor has chosen an option for which there existed a superior, or more preferred alternative. Accordingly, the analyst has to uncover the reason why the more preferred option was not chosen, which, however, leads to an ad hoc and not a systematic  John Ferejohn, "Rationality and Interpretation: Parliamentary Elections in Early Stuart England," in Kristen Monroe, ed., The Economic Approach to Politics: A Critical Reassessment of the Theory and Rational Action (New York: Harper Collins, 1991), 300301; Masaru Kohno, Japan's Postwar Party Politics (Princeton: Princeton University Press, 1997), p. 9-10. 13  4  account. In the same vein, Tsebelis claims that rational choice theory offers theoretical clarity and parsimony. Outcomes are expected to be the result of deliberate rational 14  choices rather than mistakes. This refusal of allowing for mistakes made by a rationally acting agents eliminates ad hoc explanations by forcing the analyst to refine the model and subject it to further empirical tests if the predicted outcome does not correspond with reality. Therefore, rational choice theory leads to the cumulating wisdom much more than any other paradigm where the infallibility of the theory is taken as a given. In addition, Tsebelis notes three other methodological advantages of rational choice theory: equilibrium analysis; deductive reasoning; and the interchangeability of individuals. Equilibrium analysis treats repeated political patterns as cases of equilibrium, meaning the individual actors are engaged in a situation or a relationshipfromwhich they have no incentive to depart. In other words, recurring patterns of behavior reflect the optimality of the particular course of action taken by the actor. Deductive reasoning 15  challenges the dominant, inductive, method of research in political science. The difference between inductive and deductive methods of inquiry lies in the relationship between theory and an empirical problem. Whereas inductive research proceeds from setting an empirical problem to explaining it by invoking whatever theoretical tool appears to be the most apt at providing a sound explanation, deductive research follows the exactly reverse process: first, it specifies a theoretical model and then it proceeds to applying it to an empirical problem. While the inductive method gives a wide scope of freedom to the researcher in deciding what explains the dependent variable of his/her research, the deductive method imposes strict theoretical constraints on the researcher at the very outset. According to Laver, "the essential purpose of the rational choice approach is ... to construct a logically coherent potential explanation of the phenomenon •"Tsebelis, Nested Games, p. 40.  "Ibid., pp. 41-2.  5  under investigation" , whereas the purpose of the more traditional empirical-inductive 16  approach is to "generate plausible and robust general statements that are consistent with observed empirical patterns".  17  Finally, individuals are interchangeable in rational choice accounts because their only characteristic that is assumed is rationality. Therefore, in a rational choice model it does not matter what religious, ethnic, linguistic group or gender the actor comes from because all actors seek to find an optimal mix of means to reach their goals. By allowing for the interchangeability of individuals, rational choice theory moves very close to Przeworski and Teune's requirement that a general theory in comparative politics must substitute variables for proper names.  18  While the number of its adherents has grown significantly over time, rational choice theory has also attracted a lot of criticism and challenge. The single most comprehensive criticism of rational choice to date has been offered by Green and Shapiro's earlier cited work. The summary of their position is that 19  To date, a large proportion of the theoretical conjectures of rational choice theorists have not been tested empirically. Those tests that have been undertaken have either failed on their own terms or garnered theoretical support for propositions that, on reflection, can only be characterized as banal: they do little more than restate existing knowledge in rational choice terminology. 20  In other words, the authors contend that rational choice theory is too detachedfromthe real political world; it aims only at theoretical sophistication; and in the rare case when it Michael Laver, Private Desires, Political Action: An Invitation to the Politics of Rational Choice (London: SAGE, 1997), p.4. Ibid., p.5. Adam Przeworski and Henry Teune, The Logic of Comparative Social Inquiry (New York; Wiley, 1970), p. 25. Earlier critics of rational choice theory confined themselves to articles or chapters in edited collections rather than a book-length treatment of the subject as Green and Shapiro have done. In particular, see the following works: Monroe, ed., The Economic Approach to Politics...; Jeffrey Friedman, ed., The Rational Choice Controversy: Economic Models ofPolitics Reconsidered (New Haven: Yale University Press, 1996). The latter is a collection of scholarly reactions to Green's and Shapiro's work. Green and Shapiro, Pathologies ofRational Choice, p. 6. 16  17  18  1  20  6  is applied to real empirical settings, rational choice theory offers little if any insight. Because rational choice research is theory- rather than problem-driven it fails to advance our understanding of the real political world. On the basis of this argument, Green and Shapiro identify three specific methodological pathologies of rational choice. First, they charge that rational choice theorizing has been done in an essentially post-hoc fashion. Rather than offering either new evidence about or a more plausible account of an empirical problem, practitioners of rational choice tend to simply resort to reconstructing an already known story in terms of rational choice terminology. Furthermore, rational choice, according to the authors, can hardly be called a unified theory at all, since different rational choice models often offer conflicting predictions about the same dependent variable. Second, because rational choice is preoccupied with the sophistication of theory it rarely ever offers any suggestions for the application of its models and the operationalization of its variables. In addition, rational choice works with too many unmeasurable terms, such as individual tastes and preferences, and offers predictions which are very vaguely operationalized if at all. Third, practitioners of rational choice are biased both in their selection and interpretation of evidence. In their exclusive search for theory-confirming evidence, they also tend to project the evidence itselffromthe theory thereby forcing the empirical data to fit their theoretical predictions and models. Closely related to this is what Green and Shapiro call the arbitrary domain restriction of rational choice: recognizing its inherent limitations, rational choice theory is applied to modeling only certain kinds of political behavior while entirely ignoring a range of other domains. This dissertation does not seek to do justice to either the pro- or the contra side of the rational choice divide. However, what it does seek to establish is how rational choice models and accounts fare vis-a-vis their competitors when predicting and explaining particular political phenomena. Instead of making an a priori commitment to either  7  approving or rejecting rational choice theory, this dissertation seeks to engage in a genuine test of the merits of the theory in comparison with its rival paradigms.  THE EMPIRICAL PROBLEM: PARTY SYSTEM CHANGE, GOVERNMENT FORMATION AND GOVERNMENT STABILITY IN INDIA The Indian national party system had been characterized by the dominance of the Indian National Congress Partyfromthe first post-Independence elections held in 1952, until 1989. The Congress Party, and its various incarnations born out of the multiple fissions that the organization went through, won each but one of the eight national general elections held in that period. It lost at the polls only in 1977 when a coalition of five opposition parties defeated the Congress and formed a government under the Janata Party name with a small regional party, the Akali Dal. In the next election, the Congress Party reasserted its dominance by winning the 1980 polls with a convincing margin only to increase it further in 1984. However, a fundamental transformation has taken place in the party system starting with the ninth general elections in 1989. This transformation has three important characteristics. First, in sharp contrast with the pattern of single party majorities winning the general elections, the 1989 election was thefirst,followed by the elections of 1991 and 1996, in a string of successive elections resulting in hung parliaments in which neither the Congress nor any other party managed to win a majority of the seats. The second important characteristic pertains to the increase in the number of coalition governments that political parties form. Whereas of the 10 different cabinets formed before 1989, 7 were that of a single-party majority type, and one each of the single-party minority, coalition majority and coalition minority types, the number of coalition governments in the post-1989 period has sharply increased: of the 6 post-1989 cabinets 4 were formed by coalitions of parties, although none of them commanded a majority of the parliamentary seats, and 2 were formed by single parties, both of which  8  were also in minority. The third important characteristic relates to the stability of the governments that were formed in these hung parliaments. With the sole exception of the cabinet formed by the Indian National Congress(I) Party after the 1991 election, no government managed to last for a full term in office. While there were 10 different cabinets formed in the first eight parliaments, an average of 1.25 cabinet for every legislature elected, there have been 6 different cabinets formed in the three legislatures elected since 1989, an average of 2 cabinets for every legislature elected. These changes in the Indian party system have important implications for students of both comparative and Indian politics for two reasons. The first comparative implication of the changes is that they challenge the conventional assertion that the Westminster-style parliamentary democracy, based on the single-member simple plurality electoral formula, the institutions of which India adopted, promotes the formation of stable single-party majority governments. This conventional wisdom also holds that in the rare occurrence of a hung parliament, the largest party would form a minority cabinet rather than share power in a coalition. The second comparative implication of the changes is that they are making the Indian party system more akin to the established party systems of most Western European countries where fragmented party systems, hung parliaments and coalitions governments are the normal features of party competition. This is of particular importancefromthe perspective of comparative theory given that the theories that have been formulated to explain party system fragmentation and coalition politics have been developed in the context of Western Europe making them vulnerable to accusations of an inherent socio-cultural bias. It is plausible that comparative theories of party systemfragmentationand coalition government are essentially theories designed to explain empirical observations in the Western European socio-cultural milieu. That India, a country outside the boundaries of this socio-cultural context, is providing an increasing number of observations for the  9  dependent variables of party system fragmentation and coalition government, raises the important question whether the independent variables posited in the theories designed to explain these variables in Western Europe would also account for the dependent variables in India? In other words, India does provide a good context to test the main comparative theories of party systemfragmentationand coalition politics. There are two main reasons why coalition government is an unusual phenomenon in countries that have adopted the institutionalframeworkof Britain's Westminster-style parliamentary democracy. First, this type of democracy employs the single member simple plurality electoral formula which tends to produce electoral outcomes in which one party wins an absolute majority of the parliamentary seats. Table 1.1 illustrates the election outcomes in the four largest and most established parliamentary democracies that have used this electoral system: the United Kingdom, Canada, New Zealand, and India. As the Table shows, of the 60 elections that have been held in these countries since the end of the Second World War, only 10 resulted in hung parliaments. Furthermore, it is worth noting that of these 10 instances 6 took place in Canada, 3 in India, 1 in the United Kingdom and none in New Zealand. TABLE 1.1: PARTY SYSTEM CHARACTERISTICS IN FOUR WESTMINSTER DEMOCRACIES, 19415-97  Country Canada India New Zealand Britain Total  # of elections 17 11 17 15 60  # of hung assemblies elected 6 3 0 1 10  Mean % of seats held by the largest party 55.8 59.5 57.3 54.7 56.6  Mean % of seats held by the first two parties 84.7 72.7 99.3 95.5 89.3  Second, in hung Westminster-style parliaments political parties have tended to form single party minority rather multiparty, read coalition, majority governments, see Table 1.2. Of the 65 cabinets that were formed in the 60 legislatures elected in the four  10  countries 59 were formed by a single party. Once again, the main reason for this has been the expectation, based on the tendency of the electoral formula to produce majority winners most of the time, by political parties that there will be a majority winner in the next election anyway. With this expectation in mind, there is indeed little incentive for political parties to enter into costly negotiations and build a coalition governments which might prove to be an electoral liability in the next election. Table 1.2 shows the record of the four Westminster democracies in terms of the types of government formed since the Second World War. The Table clearly illustrates that although Canada has had a noticeable record of hung parliaments, political parties did not form coalition governments in any of these legislatures. Similarly, iri the United Kingdom, although the general election of February 1974 failed to provide either of the two large parties, Conservative or Labor, with a majority of the parliamentary seats in the House of Commons, the Labor Party, which did win the most seats in that election formed a single party minority government.  21  TABLE 1.2: PARLIAMENTS AND GOVERNMENTS IN FOUR WESTMINSTER DEMOCRACIES, 1945-97  Country Canada India New Zealand Britain Total  Majority  parliament Hung parliament  Single-party majority  Coalition majority  Single-party minority  11 7 17 14 49  Coalition minority  0 1 0 0 1  Coalition majority  6 3 0 1 10  0 5 0 0 5  0 0 0 0 0  It is important to point out, however, that the idea of coalition building was not absent in Britain at the time. The outgoing Conservative Prime Minister, Mr. Edward Heath, did invite the Liberal Party to a series of negotiation about the possibility of forming an antiLabor coalition cabinet. In the end, the talks broke down, and the pattern of single-party government in Britain was maintained. 21  11  The Table reveals that India indeed deviatesfromthe other established Westminster democracies by being the only one that has seen the formation of multiparty governments in the period examined. In addition, in India's hung parliaments, political parties have formed multiparty governments more often than single party cabinets. The changes in the pattern of party competition in India have significant implications for students of Indian politics as well. First, the changes have led a number of scholars to herald the beginning of a new era in the Indian party system. For example, applying Giovanni Sartori's typology of party systems Mahendra Prasad Singh noted that the 1989 election marked a shiftfroma predominant party system to the type that Sartori called polarized pluralism. In contrast with a predominant party system featuring one 22  large party winning successive national elections with comfortable margins, polarized pluralism is characterized by afragmentedlegislature in which the level of ideological conflict among parties is very high resulting in highly unstable government coalitions. Another noted scholar of the Indian party system, Yogendra Yadav, argued that since 1989 the Indian party system has entered a phase aptly called the "post-Congress polity". Yadav argues that increased volatility andfragmentationof the party system as 23  well as the loss of the Congress Party's ability to win legislative majorities in national elections has resulted in the emergence of a distinctively new kind of party system. A main feature of the "post-Congress polity", according to Yadav, is the reflection of the multiple bipolarities of the subnational party systems in the configuration of the national party system. Multiple bipolarities refers to the presence of two main partisan poles around which party competition revolves in each state of the country. Since these bipolarities are regionally specific with different parties representing the poles in the different regions of the country, the aggregation of these subnational patterns at the Mahendra Prasad Singh, "From Predominance to Polarized Pluralism: The Indian Party System," Think India 2 (1990), pp. 70-8. Yogendra Yadav, "Reconfiguration in Indian Politics: State Assembly Elections, 1993-  22  23  95," Economic andPolitical  Weekly 31 (1996), pp. 94-105.  12  national level yields a highly fragmented picture. It is worth noting, however, that while Yadav claims this system of multiple bipolarities to be the hallmark of the post-Congress polity, the Congress Party does remain one of the local poles in each region and state of India with the exception of the large states of Uttar Pradesh and Bihar in the North. This suggests that the Congress still continues to be an important electoral force to reckoned with even thought its dominance may now be a matter of the past. While the changes captured by the terms polarized pluralism and post-Congress polity highlight what is new and different in the party system, they overlook some important continuities in the logic and dynamics of party competition. For example, while the election of hung parliaments is certainly a novel phenomenon at the national level, the lower house of the national parliament was hung before as a result of splits in the political party that had won a majority of seats in the election. This had happened in the national legislature twice prior to 1989: for the first time in 1969 when the governing Indian National Congress Party was split between two rival factions, and for the second time in 1979 when the governing Janata Party was split between its two factions. In addition, such splits have also occurred on numerous occasion before at the subnational level. This suggests that political actors had already had experience in forming and maintaining governments in hung parliaments before the election of such assemblies became a recurring pattern of the party system as of 1989. Another important continuity easily overlooked by focusing on the decline of the Congress Party and the transformation of the predominant party system pertains to the continued importance of the role of the Congress(I) Party itself in national politics. As Chapter 5 will elaborate in more detail, by its position as a center party, the Congress continues to occupy the pivotal position of determining government stability in the current party system. In other words, so long as the Congress remains a center party and the national legislature remains divided, no cabinet can survive for long without the Congress Party's explicit or implicit support. Conversely, the Congress remains in a very  13  strong position to form a stable government because its center position allows it to divide the opposition which thus cannot overthrow it. In sum, although the Congress is no longer a dominant party in terms of its numerical strength in the national legislature, it still remains the pivot of the party system around which the system's dynamics revolve. As such, it is perhaps a bit too early to talk about a post-Congress polity at this juncture. Given the comparative and India-specific implications of the changes outlined above, the dissertation will seek to provide an answer to three questions: 1. Why has the predominant party system been replaced with a fragmented party system featuring hung parliaments since 1989? 2. Why have political parties formed single-party minority, coalition minority and coalition majority governments in India's hung national parliaments? 3. Why did some governments that political parties formed in India remain stable while others have failed to last for their full term in office?  METHODOLOGY The design of this study seeks to eliminate the two methodological problems that are often encountered in comparative politics: the problem of too many variables and too few cases, noted by Lijphart , and the problem of selection bias noted by Geddes . 24  25  According to Lijphart the scientific study of comparative politics is plagued with the problem of too many variables and too few cases. This means that when the number of cases to be compared is too few and the cases are too different on a range of background variables, it becomes very difficult to provide a robust comparative explanation for the dependent variable in question. The simple reason is that the number  Arend Lijphart, "Comparative Politics and the Comparative Method", American Political Science Review 65 (1971), pp. 682-93. "Barbara Geddes, "How the Cases You Choose Affect the Answers You Get: Selection Bias in Comparative Politics" in James A. Stimson, ed., Political Analysis volume 2. (Ann Arbor: University of Michigan Press, 1990) 24  14  of possible explanations for the dependent variable increases with every background variable on which the cases differ. Collier also notes that as "the number of explanatory factors approaches the number of cases, the capacity to adjudicate among the explanations through statistical comparison rapidly diminishes."  26  Lijphart's solutions to the problem of many variables and small N included calls for increasing the number of N as well as for the pursuit of the comparable cases strategy, i.e. a research strategy that rests on the selection of cases that are both similar on most variables that are not central to the study and different in terms of the key variables that the analysis focuses on. In this, Lijpahrt advocates a highly controlled method of comparison similarly to Eggan and Stinchcombe . In contrast, Przeworski and Teune 27  28  29  have argued that pursuing highly controlled comparison, in what they called the strategy of most similar systems, may lead to the problem of overdetermination. Przeworski and Teune were concerned that too much control in the selection of cases may result in the choice of cases which are so similar that mostrivalexplanations will explain the dependent variable, thus leaving the researcher with no clue to decide which explanation is the best. The solution that they suggest is to pursue a "most different" cases strategy whereby the cases to be compared are very different in terms of a wide range of background variables but are similar on the dependent variable. This approach was challenged by Geddes who argued that if the cases, no matter how different, are chosen 30  David Collier, "The Comparative Method" in Ada W. Finifter, ed., Political Science: The State of the Discipline II. (American Political Science Association, 1993), pp. 105119. Fred Eggan, "Social Anthropology and the Method of Controlled Comparison", American Anthropologist 56 (1954), pp. 743-63. Arthur Stinchcombe, Theoretical Methods in Social History (New York: Academic Press, 1978). Adam Przeworski and Henry Teune, The Logic of Comparative Social Inquiry (New York: John Wiley, 1970). Also see, Adam Przeworski, "Methods of Cross-National Research, 1970-1983: An Overview" in Meinolf Dierkes, Hans N. Weiler, and Ariane 26  27  28  29  Berthoin Antal, eds., Comparative Policy Research: Learningfrom Experience (Brookfield: Gower, 1987). Geddes, "How the Cases You Choose..."  30  15  solely on the basis of their similarity on the dependent variable, the research suffers from a selection bias which reduces the power of the explanation. Geddes suggested that one way in which such a bias may be avoided is by increasing the variation on the dependent variable, for example through internal comparison of the same case, such as an intertemporal comparison of observations within the same system. In this study, both sides of the "too many variables and small N" problem will be alleviated. The problem of too many variables is reduced by controlling for a significant range of background variables of each observation given that the cases are selected from the national party system of one and the same country, India. The other side of the problem, i.e. small N, is alleviated as much as possible by extending the time horizon of the study over the entire history of Independent India's party politics. By so doing, all instances of each dependent variable are examined, resulting not only in a larger N than what a study focusing solely on the post-1989 scene would yield but also variation in the dependent variable. For example, an explanation that can account for the formation of governments of all types in India is certainly more powerful than one that could consistently explain the formation of the post-1989 coalitions but would fail to account for the pre-1989 instances. With these methodological considerations in mind, the ensuing chapters will be structured in the following way: •  Chapter Two will provide the empirical information about the history of the Indian party system, government formation and stability in Independent India.  •  Chapter Three will examine the transformation of the Indian party system from predominant to bipartism in 1977,frombipartism back to predominance in 1980 and from predominance to an even multiparty system in 1989. The chapter will especially seek to contrast the inability of the Congress Party to make an electoral comeback since its loss in 1989 with the re-equilibration of its electoral dominance in 1980 following its defeat in 1977.  16  •  Chapter Four will seek to account for the formation of different types of governments in India: single party majority by the Congress Partyfrom1952 through 1969, in 1971, and again in 1980 and 1984; single party minority by the Congress (Ruling) in 1969, the Samajwadi Janata Party in 1990, and the Congress (I) in 1991; coalition minority by the Janata Party (Secular) and the Congress in 1979, the National Front in 1989, the Bharatiya Janata Party anddhe Shiv Sena in 1996, the United Front (Gowda) in 1996, and the United Front (Gujral) in 1997; and coalition majority by the Janata Party and the Akali Dal in 1977.  •  Chapter Five will seek to contrast and explain the differential durability of the various Indian governments. The variation in the dependent variable, government durability, in this chapter is provided by the contrast between the stability of the Congress majority governments of 1952, 1957, 1962, 1971, 1980, 1984 and 1991 and the failure of all other governments to last their full term in office.  •  Chapter Six will attempt to weave together the three analytical themes of the earlier chapters. By drawing on George Tsebelis' theory of nested games, it will offer a rational choice model in which the three themes can be united.  COMPARATIVE STUDIES ON THE INDIAN PARTY SYSTEM AND COALITION POLITICS An important reason that makes a comparative study of India very important pertains to the divergence in the evolution of the comparative party politics and the Indian party politics literatures. To put it plainly, while comparativists have paid only scant attention to the applicability of their models and theories to the Indian context, India specialists have likewise pursued their inquiries in isolationfromthe research agenda of the comparative party politics literature. Notwithstanding this general divergence of the two literatures, there have also been four noteworthy attempts to bridge the gap between them. Thefirstand classic  17  exception among comparativists is provided by William Riker. Although his account will be examined in more detail in Chapter 3, it is important to stress here his awareness of the importance of Indian exceptionalism in the study of party system fragmentation. In his 1976 article Riker proposed an ingenious theory on the number of political parties in single member simple plurality electoral systems. According to Riker, it is the presence/absence of an underlying ideological dimension and the location of both voters, who are assumed to be rational and to vote either strategically or sincerely, and parties along this dimension, that determine the number of parties. Riker reviews a number of possible scenarios and finds that it is possible for the party system to reach equilibrium with one dominant and many tiny parties under the single member simple plurality formula, provided that there is an underlying ideological dimension and that the dominant party occupies the center of it. Riker cites India as the quintessential empirical example of such a case noting that although India employed the single member simple plurality formula, instead of a twoparty system there emerged a dominant party system featuring the Indian National Congress Party as the dominant actor surrounded by a plethora of small parties in both the electoral and the legislative party systems. In a later review of the literature on the development of political science since Duverger's Law, Riker re-emphasized the importance of Indian exceptionalism. In sum, Riker's contribution not only helped to clarify and refine Duverger's Law, by highlighting the case of India it also provided a truly comparative theory of the emergence of one-party dominance in a competitive multiparty democracy. While Riker's mentioned works have become classics in the literature on comparative party systems, comparativists have failed to recognize that the Indian party system has since moved on from one party dominance to becoming a genuine multiparty democracy thus providing a different kind of exception to Duverger's Law. Whereas prior to 1989 the Indian party system did not have enough large parties to be classified as  18  a two-party system, since 1989 the party system has become sofragmentedthat the number of significant parties has exceeded two. While this transformation of the party system has led India specialist scholars to herald the beginning of a new era in Indian politics, comparativists have as yet failed to appreciate the theoretical significance of this change. The other three works that have attempted to link studies of the Indian party system with the comparative literature are provided by Paul Brass , Subrata Kumar 31  Mitra and Bruce Bueno de Mesquita who have tried to account for the formation and 32  33  instability of the coalition governments that were formed in India at the subnational level. Paul Brass made an explicit attempt to statistically test a set of hypotheses derived from a study on government stability in Western Europe by Taylor and Herman in the 34  context of the Indian states. He found that most hypotheses put forth by Taylor and Herman were supported by the Indian data, "but in all cases at lower levels of explanation than in their study" (1404). Specifically, Brass found it confirmed that just as in Western Europe, one-party majority governments tend to outlast coalition governments in the Indian states as well. However, Brass also found that unlike in Western Europe, the standard deviationfromthe mean degree of stability within each type of government is significantly greater in the Indian states than in Western Europe (1389). This implies that Indian state government follow the predicted pattern of stability much less consistently than Western European national governments.  Paul Brass, 'Party Systems and Government Stability in the Indian States', American Political Science Review 71 (1977), pp. 1384-1405. Subrata Kumar Mitra, Governmental Instability in Indian States: West Bengal, Uttar Pradesh and Punjab (Delhi: Ajanta Publications, 1978). Bruce Bueno de Mesquita, Strategy, Risk and Personality in Coalition Politics: Th Case of India (New York: Cambridge University Press, 1975) 31  32  33  ^Michael Taylor and Valentine Herman, "Party Systems and Government Stability," 28-37.  American Political Science Review 65 (1971), pp.  19  Bruce Bueno de Mesquita's work on subnational coalitions in India goes a long way towards bringing India and comparative politics closer to one another. Mesquita set out to go beyond traditional works on Indian party politics in two senses. First, whereas previous studies on party politics had either focused on all parties in one selected state of India or one political party across all or most states, Mesquita analyzes the interaction among several parties in several states in the processes of coalition building, maintenance and termination and the electoral performance of coalition members. Second, Mesquita 35  also consciously wants to provide a comparative explanation for his observations. As he claims, his theory suggests an explanation ...without exploring the very complex patterns of caste, regional and linguistic influences on voter preferences. Instead, the exchange of legislative seats is accounted for with considerable precision, from knowledge of the strategic preferences of decision-makers, the organizational capabilities of their parties, and differences in the level of need for achievement possessed by the leaders of each party. 36  Thus, Mesquita explicitly sets out to avoid explaining away his dependent variables in terms of unfalsifiable idiosyncratic variables. However, at the same time he fails to compare his explanation with competing rival accounts of coalition politics offered either in the Indianist or the comparative literatures. Thus, no matter how insightful his analysis the reader is left uncertain whether his explanation is indeed superior to what has already been offered in other literatures. A similar rational choice approach to coalition politics at the subnational level is offered by Subrata Kumar Mitra. Mitra found that the [stability of the formation and maintenance of winning coalitions, particularly in terms of the spatial location of the position of the cabinet have been shown to be a the magnitude of factionalism within the winning coalition. It was shown that a stable outcome equivalent to cabinet stability exists when all the players use one overall dimension (i.e. an ideological dimension) rather than issues of individual Bruce Bueno de Mesquita, Strategy, Risk and Personality, p. 2.  35  Ibid.  36  20  concern (47). While retaining the importance of the impact of ideological dimensions on government formation and stability, which has been established in the comparative literature, Mitra also went a step further by systematically including the additional variable of factionalism in his analysis. In so doing, Mitra actually preceded comparativists who have only recently started looking at the importance of this variable.  37  The model that emergesfromMitra's analysis constitutes a milestone in the literature on Indian party politics even though it has been largely marginalized. The model that he constructed to explain government stability combines elements of two distinct schools of thought in the comparative literature: the so-called 'attribute' and spatial schools. In brief, he argues that the probability of a complete structuration of the electorate, with a completely structured electorate meaning that all voters position themselves along a single dimension of conflict, is negatively related to the fragmentation of the parliamentary party system, which, according to the "attribute" school of thought is positively related to cabinet instability. At the same time, the probability of a completely structured electorate also increases the probability that the legislative party system will also be unidimensional and factionalism reduced to the minimum, which is negatively related to cabinet instability, according to the spatial school of thought. In sum, a completely structured unidimensional electorate will promote cabinet stability, according Mitra's model. Apartfromthese attempts, the current state of scholarship on party politics in India remains within the tradition of area-studies offering little connection with the theoretical advances made in the comparative literature.  See Michael Laver and Kenneth A. Shepsle, "Government Coalitions and Intraparty Politics", British Journal of Political Science 20 (1990), pp. 489-507. 37  21  CHAPTER TWO: PARTY POLITICS IN INDEPENDENT INDIA 2.1  THE FIRST THREE LOK SABHAS, 1952-1967: THE DOMINANCE OF THE CONGRESS PARTY The first three post-Independence elections resulted in massive victories by the  Indian National Congress Party. Although the Congress party did not win a majority of the popular vote cast in any of these three elections, it consistently won a majority of the seats in the Lok Sabha. According to parliamentary convention, the leader of the Congress Parliamentary Party, in each instance Jawaharlal Nehru, was invited by the President of the Republic to form a government in each of these Lok Sabhas. The preeminent position of the Congress was clearly reflected in the difference between the share of the popular vote won by the Congress and the second largest party: in the 1952 elections the difference was 43.4%, in 1957 it was 37.4%, and in 1962 it was still very high at 34.8%. In terms of legislative size, the dominance of the Congress was even more starkly reflected: in the first Lok Sabha the Congress had 71.2% more seats than the second largest party in the Lok Sabha, in 1957 it had 69.6%, and in 1962 it had 67.2% more seats than the largest opposition party. The Congress Party enjoyed a similar degree of dominance at the subnational level as well. Of the 42 state elections, all but two of which were held concurrently with the national polls, during this period only four resulted in the formation of non-Congress governments: in 1957 the Communist Party of India, in 1960 the Praja Socialist Party formed governments in the state of Kerala, in 1963 the Maharastrawadi Gomantak party formed a government in Goa, and in 1964 the Naga Nationalist Organization formed a government in Nagaland. Thus, the first decade and a half of India's post-Independence party politics was characterized by the weakness of intense inter-party competition. Given the Congress Party's dominance over both the electoral and the legislative spheres Kothari aptly  22  labeled it the Congress system. According to Kothari's analysis, the role of the 38  opposition parties in this system was not to constitute and offer a real alternative to the Congress government but rather to ...constantly pressurize, criticize, censure and influence it by influencing opinion and interests inside the margin and, above all, exert a latent threat that if the ruling group strays away too far from the balance of effective public opinion, and if the factional system within it is not mobilized to restore the balance, it will be displaced from power by the opposition groups. 39  In turn, what allowed the Congress to respond to these pressures and remain in power for so long without interruption was its internal plurality,flexibility,internal competitiveness and its capacity to absorb rival movements thus preventing the opposition parties from growing.  40  Kothari's analysis of the working of the Congress system is problematic because it does not clarify why the opposition parties would want to keep criticizing and pressurizing the dominant party in such a way that would allow the latter to remain in power at their own expense! Intuitively it would appear to be more logical for the opposition parties to actually work towards pressuring the dominant party to move away from the "effective balance of public opinion" so that they themselves could enter into office. Nonetheless, his model offers a good description of the early Indian party system that was indeed dominated by the Congress Party. In the period of the Congress system, the formation of single-party majority governments was the norm as the Congress Party consistently won legislative majorities in the national elections. In this period, coalition politics at the national level was non-existent. Rajni Kothari, "The Congress 'System' in India," in Party System and Election Studies, Occasional Papers of the Center for Developing Societies, No. 1 (Bombay: Allied Publishers, 1967), pp. 1-18. A similar characterization of the party system of this period was offered by W. H. Morris Jones, "Parliament and Dominant Party: India Experience," in Parliamentary Affairs 17 (Summer, 1964), pp. 296-307. Kothari, "The Congress 'System'," p. 3. 38  39  lbid, p.6.  m  23  2.2  T H E F O U R T H L O K S A B H A , 1967-71: SPLIT IN THE G O V E R N I N G P A R T Y A N D THE FIRST MINORITY G O V E R N M E N T  41  The general election of 1967 was an important landmark in India's electoral history for two reasons. First, the Congress Party was badly weakened in the election both in terms of the popular vote it mustered as well as the legislative seats that it won. For the first time since Independence, the Congress Party barely managed to win more than 40% of the total national vote. In addition, for the first time it failed to secure a twothirds majority in the Lok Sabha: by winning 54% if the seats in the house the Congress still enjoyed a thin majority, however, its size paled in comparison with the party's strength in the earlier parliaments. Second, the dominance of the Congress Party was also shattered in the 1967 election at the subnational level. The Party failed to win a majority of the seats in the State Assemblies of Bihar, Kerala, Orissa, Punjab, Rajasthan, Tamil Nadu, Uttar Pradesh and West Bengal. Except for Rajasthan where civil violence erupted in the aftermath of the elections and President's Rule was imposed by the national government, the nonCongress parties managed to exploit the weakness of the Congress and formed governments in these states. In short, the 1967 election left the Congress in a badly weakened state at the national level and out of power in some of the largest states of the country. The Congress Party had also been suffering with internal problems since the death of its long-time leader and Prime Minister, Jawaharlal Nehru, on May 27, 1964. After 42  Nehru's death a group of state-level party bosses, called the Syndicate, united behind K . Kamaraj, president of the Congress Party, to ensure the election of Lai Bahadur Shastri as  ( 1 9 7 ^ Pp! 2 3767- 9 h  e  U  SSi0n  " ^  S C C t i  ° " ^ n  ^  C  o  n  t  e  m  P  o  r  ^  Archives  -Robert L. Hardgrave and Stanley A . Kochanek, India: Government and Politics in a Developing Nation (Fort Worth: Harcourt College Publishers, 2000), p. 251.  24  Nehru's successor in the Prime Ministerial berth. Following Shastri's death after only 43  two years in office, the Syndicate also engineered the succession of Nehru's daughter, Indira Gandhi, following a bitter contest with Morarji Desai, the Congress Party's strong man from the state of Gujarat. The 1967 election weakened the position of the Syndicate very badly. Kamaraj himself as well as nine ministers in the national cabinet and four other Chief Ministers, the equivalents of the Prime Minister at the state level, failed to retain their seats. The weakening of the Syndicate, however, strengthened the position of both Indira Gandhi, who had been seeking to distance herself from the influence of the bosses, and Morarji Desai, who never gave up on his ambition to become the Prime Minister. The apparently inevitable confrontation among the three players led to the split of the Congress Party and the government losing its majority in the Lok Sabha. The series of events that led to the open split between the Syndicate and Indira Gandhi started with the All-India Congress Committee session held in Bangalore on July 10, 1969. The night before the session, the Congress Working Committee held a 44  meeting to draft a resolution on the Congress government's economic policy to be presented for ratification to the session the next day. The Industrial Development Minister, Mr. Fakhruddin A l i Ahmed, delivered the Prime Minister's memorandum on economic policy, as Indira Gandhi was unable to attend the pre-session meeting. The Prime Minister's memorandum put forward the following proposals: 0 n the grand succession struggles in the Congress party, see Michael Brecher, Nehru's Mantle: The Politics of Succession in India (New York: Praeger, 1966). The All-India Congress Committee represents a layer in the decision-making structure of the Congress Party. The Committee is made up of one eighth of each state's delegation to the Annual Congress Session and elected by only the delegates from the given state. The All-India Congress Committee elects seven members of the Congress Working Committee, the highest executive organ of the party. The Working Committee also includes the Congress President who appoints thirteen members of the Committee at his/her discretion. The President of the Congress Party is elected by all delegates to the Annual Congress Session for a two-year term. See, Stanley A . Kochanek, The Congress Party of India (Princeton, N . J.: Princeton University Press, 1968), p. xxii. 43  44  25  •  a ceiling on unproductive expenditure and conspicuous consumption by corporate bodies;  •  nationalization of private commercial banks and a change in their credit policy to make it more favorable to new entrepreneurs, especially in less developed regions;  •  expeditious appointment of a monopolies commission, more autonomy and professionalization of the public-sector enterprises, reserving consumer industries for the small-scale private sector and banning the entry of big business into this field;  •  providing more avenues of employment to the young and educated;  •  disallowing foreign capital in areas where local technical know-how was available;  •  special assistance to small agriculturists and farm cooperatives, and the legislation and implementation of a more vigorous scheme of land reforms aimed at protecting the rights of tenants, distribution of surplus land yielded by placing ceilings on land holdings and of government wastelands among the landless, prevention of the alienation of landholdings belonging to the scheduled castes and the scheduled tribes, and a wage review for agricultural laborers to give them a share in the increased agricultural yields in areas where the "green revolution" had made some headway.  45  This memorandum constituted a radical departure from the policy that the A l l India Congress Committee had adopted the year before, in 1968, at Jabalpur. In particular, what made it radically different was the inclusion of the clause calling for the nationalization of banks. Apart from its ideological implications, bank nationalization was a critical issue also because banking regulation belonged to the jurisdiction of the national government. In contrast, the other provisions, as well as those of the Jabalpur resolution, were about matters that were under state jurisdiction and thus they allowed the Syndicate-controlled state Congress units to exercise considerable discretion in changing ^ ! ^ " ^ ' ° P^ty: The Indian National Congress in 1969 (New Delhi: Abhinav Publications, 1981), p.p. 63-4.  T  1  1  S  I  N  G  H  S  p  l  U  i n  P  r  e  d  o  m  i  n  a  n  t  26  and moderating the policy resolution of the national party organization at both the legislative and the policy implementation stages. In other words, Indira Gandhi's call 46  for bank nationalization sought to give greater control to the national party organization over party policy, and since the party was also in office at the national level, to the national government over public policy. Although formally the Working Committee unanimously accepted Indira Gandhi's proposals, the Syndicate joined hands with its former archenemy, Morarji Desai, in criticizing the plan on bank nationalization. Desai, who was both Finance and Deputy Prime Minister at the time warned that he would rather resign than give his name to the implementation of the plan. Nonetheless, indicative of the power of Indira 47  Gandhi's group, the Working Committee got Desai to actually move the resolution to the All-India Congress Committee. Apart from economic policy, nominating the party's official candidate for the looming Presidential election was another prominent issue that the Bangalore session had to deal with. The Syndicate's favorite for nomination was one of its own members: the Speaker of the Lok Sabha, Mr. Sanjiva Reddy. In contrast, the Prime Minister and her supporters preferred the nomination of either Acting President V . V . Giri, or Jagjivan 48  Ram, the Food and Agricultural Minister, or Swaran Singh, the Defense Minister, for the post. Although the majority of the Working Committee preferred Sanjiva Reddy's nomination the Prime Minister issued a warning that such a sensitive issue should be settled by consensus rather than majoritarian decision-making. Eventually, as the two sides failed to reach a compromise, Indira Gandhi suggested to the new president of the Congress Party, Mr. Nijalingappa, that Reddy's nomination by the Congress Party should '"Ibid., p. 64. Singh, Split in a Predominant Party, p. 64-5. In the 1967 Presidential election Zakir Hussain was elected President and V V Giri  47  48  X * W of P President A \ Hussain u death inTMay 1969. a  S  e  d  t h e  P O S i t i  °  n  °  f A c t i n  §  P r e s i d e n t  a f t e r  the untimely  27  be made public. Shortly after the conclusion of the Bangalore session, Indira Gandhi announced on July 16, 1969 that she removed Morarji Desai from the Finance portfolio which would now be under her own direct charge. According to Mrs. Gandhi, the reason behind the removal of Mr. Desai was his apparent lack of commitment to implementing the economic policy resolutions arrived at Bangalore, most importantly the policy on bank nationalization. In reaction, Mr. Desai resigned from his post as Deputy Premier as well.  49  On July 19, just a week before the parliamentary session was to start, the Prime Minister advised acting President Giri to promulgate the Banking Companies (Acquisition and Transfer of Undertakings) Ordinance, which effectively nationalized 14 of the major banks of the country. The opposition parties were divided in their reaction to the ordinance. The rightwing Swatantra Party and the Bharatiya Jana Sangh were clearly against it whereas two farmers' parties, the Bharatiya Kranti Dal and the Akali Dal, and also the Dravida Munnetra Kazagham, the party advocating the interests of the state of Tamil Nadu, and all the left parties including the Communist Party of India, the Communist Party of India (Marxist), the Samyukta Socialist Party and the Praja Socialist Party, expressed their unreserved approval. The conflict over bank nationalization was hardly over when the Congress had to deal with the other divisive issue, namely the Presidential nomination, as election to Rasthrapati Bhavan, the Presidential Palace, was scheduled for August 16. The President of the Republic in India is elected by an electoral college using a somewhat complex formula that involves both the national and the state legislatures. This was of particular 50  "Ibid., pp. 66-7.  According to Article 54 of the Indian Constitution, The President shall be elected by the members of an electoral college consisting of (a) the elected members of both Houses of Parliament; and (b) the elected members of the Legislative Assemblies of the states. 50  28  importance at the time since the results of the 1967 national and state elections left the Congress Party in a significantly weakened position in the Presidential electoral college. As Table 2.1 shows, the Congress Party did not actually have a majority therein. This meant that the enforcement of party discipline was imperative for both the Congress and the opposition to ensure the election of their respective Presidential candidates. T A B L E 2.1: THE N U M B E R OF VOTES C O N T R O L L E D B Y THE CONGRESS P A R T Y A N D THE  Lok Sabha Rajya Sabha Vidhan Sabhas College total  Congress Party 296946  Opposition 233064  235771 532717  302396 535460  Decides, various tables.  Furthermore, Article 55 sets out the actual method of selecting the President. (l)as far as practicable, there shall be uniformity in the scale of representation of the different States at the election of the President. (2) For the purpose of securing such uniformity among the States inter se as well as parity between the States as a whole and the Union, the number of votes which each elected member of Parliament and of the Legislative Assembly of each State is entitled to cast at such election shall be determined in the following manner; (a) every elected member of the Legislative Assembly of a State shall have as many votes as there are multiples of 1000 in the quotient obtained by dividing the population of the State by the total number of the elected members of the Assembly; (b) if, after taking the said multiples of one thousand, the remainder is not less than five hundred then the vote of each member referred to in sub-clause (a) shall be further increased by one; (c) each elected member of either House of Parliament shall have such number of votes as may be obtained by dividing the total number of votes assigned to the members of the Legislative Assemblies of the States under subclause (a) & (b) by the total number of the elected members of both Houses of Parliament fractions, exceeding one half being counted as one and other fractions being disregarded. The election of the President shall be held in accordance with the system of proportional representation by means of the single transferable vote and the voting at such election shall be by secret ballot.  29  The debate within the Congress Party about the issue of Presidential candidate nomination accelerated when on August 11 two backbenchers of the party, Arjun Arora and Shashi Bhusan, resigned in protest against Nijalingappa's appeal to the opposition Jana Sangh and Swatantra parties to support Reddy's candidacy. The Prime Minister supported the two backbenchers in their demand that there be no party whip imposed on the Presidential vote and that legislators could vote according to their conscience instead. Indira Gandhi's faction was particularly worried that a candidate backed by the Syndicate and the right-wing opposition might occupy Rashtrapati Bhavan, the Presidential palace, because these parties favored a constitutional arrangement that would give greater powers to the President.  51  Moreover, although the Indian President's powers are highly circumscribed and are largely ceremonial as long as there is a solid majority party government in office in the Lok Sabha, in case of a hung parliament the President's position becomes much more important as he/she can decide whom to invite to form a government. Since the Prime Minister's position in the Lok Sabha was becoming increasingly tenuous as tensions within her governing party were mounting, it was very important for her that Rashtrapati Bhavan would be filled by her supporter. The Presidential contest came down to a race between two candidates: Sanjiva Reddy, the official nominee of the Congress Party who was also supported by the rightwing parties, and V . V . Giri, running as an Independent and supported by those Congress legislators who were in favor of the conscience vote as well as the left-of-center parties. At the state level, only in Jammu and Kashmir, Punjab and Himachal Pradesh were the 52  state units of the Congress Party in favor of Giri's candidacy. In each of these states the The Swatantra and the Jana Sangh had already indicated their preference for a stronger Presidency when they had petitioned Giri to withhold the Bank Nationalization Bill. See, Singh, Split in a Predominant Party..., p. 73. Although the leader of the Congress legislature Party in Punjab, Harinder Singh, floated a "vote Reddy" circular to his party's representatives he refrained from issuing a formal whip. Singh, Split in a Predominant Party, p. 78. 51  52  30  Chief Minister/Congress Legislature Party Leader as well as the president of the state party organization opted for free conscience voting in the presidential election. In the states of Bihar, Uttar Pradesh, Orissa and West Bengal, the leader of the Congress Legislature Party and the leader of the party organization took conflicting positions on the issue. In the remaining states, both the Congress Legislature party and the party organization were in favor of a whipped vote, which meant a vote for the Syndicatebacked Sanjiva Reddy.  33  On August 16, V . V . Giri won the Presidential election by securing 50.2% of the votes against Sanjiva Reddy's 48.5%. On September 19, Congress MPs close to the 54  Syndicate filed a petition in the Supreme Court challenging the outcome of the Presidential election. Furthermore, Morarji Desai made public announcements about the possibility of overthrowing Indira Gandhi's cabinet. In turn, the Prime Minister removed four pro-Syndicate members of her cabinet as well as started a signature campaign in order to convene a special session of the All-India Congress Committee that would have the authority to suspend Nijalingappa's term as party president.  55  The Congress Working Committee was going to hold a meeting on November 1 to address the issue of party unity. However, the day before the meeting Nijalingappa had withdrawn the invitation to three pro-Indira Gandhi members of the Committee to attend. These members along with the Prime Minister held a parallel Working Committee meeting at her residence where it was resolved that the special session of the All-India Congress Committee be convened. On November 12, the Syndicate-led Working Committee officially expelled Indira Gandhi from the Congress Party and instructed the Congress Parliamentary Party, consisting of all Congress members in both the Lok Sabha and the Rajya Sabha, to choose a new leader. In the meantime the pro-Indira Working 5 3  M . P. Singh, Split in a Predominant  Party, p. 76.  Butler et al., India Decides p. 60. Note that Giri secured this majority only on the second ballot as the first preference votes did not yield a majority winner. 54  "Singh, Split in a Predominant  Party, pp. 88-91.  31  Committee resolved that Indira Gandhi's expulsion was not only illegal as per the rules and regulations of the Congress organization and the Congress Parliamentary Party, but it is also a disastrous move with far-reaching political and constitutional implications.... Mrs. Indira Gandhi continues to be the leader of the party unless the party expresses no confidence of the leader. As long as she continues to enjoy the support of parliament, she is entitled to function as the Prime Minister. It will be the saddest day for Indian democracy i f a small group of men, without even having the mandate of the organization for this purpose, should dare take a step with a view to displacing the Prime Minister. This is against all canons of parliamentary democracy. It is our considered view that these gentlemen have no legal or political authority to remove the Prime Minister of the country ... 56  The Congress Parliamentary Party met on November 13 and passed a vote of confidence in Indira Gandhi, however, only 330 of the total 427 members attended the meeting. It was not surprising at all to see the Congress Parliamentary Party side with the incumbent Prime Minister because both the bank nationalization and the Presidential nomination battles clearly showed that Indira Gandhi was stronger than the Syndicate. Congress representatives decided to stay with Indira Gandhi simply because she was more likely to retain her hold on power and be able to reward her supporters. Three days later, on November 16, the Congress Parliamentary Party officially split thus depriving the incumbent government of its majority. On that day, 60 Lok Sabha Congress representatives and 36 Rayja Sabha Congress representatives met and elected Ram Subhag Singh as the party's leader in the Lok Sabha, S. N . Mishra as the party's leader in the Rajya Sabha, and Morarji Desai as the chair of the Congress Parliamentary Party. This group came to be called the Congress Party (Opposition), although soon thereafter the name was changed to Organization Congress. The pro-Indira Gandhi group retained the Congress party label but it soon came to be referred to as the Congress Party (Ruling). The Congress (Ruling) initially had the support of 210 Congress representatives in the Lok Sabha, which soon swelled to 221 and 92 members in the Rajya Sabha respectively. Keesing's Contemporary Archives (1970), p. 23767.  56  32  T A B L E 2.2: THE C H A N G I N G POSITION OF THE CONGRESS P A R T Y IN THE F O U R T H L O K S A B H A , 1967-71  Party Congress Congress (Ruling) Congress (Opposition)  # of seats after the election 283 NA NA  % of seats aften # of seats after the elections Congress split 54.90% NA NA 221 NA 64  % of seats after Congress split NA 42.70% 12.40%  Note: The United Independent Group consisted of the Akali Dal, the Moslem League, and the Revolutionary Socialist Party.  The opposition parties welcomed the split in the Congress. The right-wing Bharatiya Jana Sangh and Swatantra Party both looked forward to the possibility of forming a new coalition government of the democratic, read non-Communist, opposition parties. The Praja Socialist Party pledged its conditional support to the government "judging every issue on its merits" and called on Indira Gandhi to consolidate the 57  democratic socialist forces of the country. The Communist Party of India, CPI, also offered its conditional issue-based support to the government and called for unity of the left parties in Parliament. In a statement issued by the party on November 26, it explicitly identified the coalition of the Syndicate, i.e. the Organization Congress, the Bharatiya Jana Sangh and the Swatantra Party as the main enemy to fight. The Communist Party of India (Marxist) also committed itself to providing issue-based support for the minority government and stated that it would certainly not vote in favor of a no confidence motion against the Prime Minister. The party further stated that its support for the government would depend on the latter's implementation of specific policies, such as the normalization of relations with China, reduction of the country's dependence on foreign aid, and progressive economic and political reforms. The only party that was highly divided over the issue of extending support to the Congress (Ruling) was the Samyukta Socialist Party. The moderate faction of the party called for adopting a position of general  "Keesing's Contemporary Archives (1970), p. 23768.  support and specific issue-based opposition whereas the more radical faction, led by Madhu Limaye, insisted that the party should continue its position of "general opposition and specific support."  58  On November 17, Piloo Moody of the Swatantra Party moved a motion of noconfidence in the government. The motion was rejected by 306 against and 140 in favor. Those voting in favor included 57 Congress (O) members, the BJS, the Swatantra Party, the Praja Socialist and the Samyukta Socialist Parties and some Independents. The Congress(R)'s 209 members, the Communist Party of India, the Communist Party of India (Marxist), the Dravida Munnetra Kazagham and some Independents voted against the motion. The Congress(R) government could have continued in office as long as it was not defeated on a vote of confidence on the floor of the Lok Sabha. Although the minority Congress(R) government was successful in maintaining the legislative coalition that allowed it to continue in office, it chose to dissolve the Lok Sabha in December 1970 and call new elections to be held in the spring of 1971, a year before they would have been due. The chain of events leading to the dissolution of the Lok Sabha began on May 18, 1970 when Home Minister Y. B. Chavan introduced a constitutional amendment bill in the lower house seeking to replace Articles 291, 362 and Clause 22 of Article 366. These Articles defined the relationship between the Government of India and the former princely rulers with whom the Government of India entered into a covenant after Independence.  59  Keesing's  5S  Contemporary  Archives (1970), p.23768.  The Articles contained the following the provisions: Article 291. Where under any covenant or agreement entered into by the Ruler of any Indian State before the commencement of this Constitution the payment of any sums, free of tax, has been guaranteed or assured by the Government of Dominion of India to any Ruler of such State as privy purse, (a) such sums shall be charged on, and paid out of, the Consolidated Fund of India, and (b) the sums so paid shall be exempt 59  34  Although the decision to abolish the privy purses was taken by the Congress Party prior to its split, the leader of the Congress (Opposition), Morarji Desai strongly opposed this constitutional amendment bill. He agreed that privy purses should be abolished; however, he also demanded that adequate compensation be made to the princes. In addition, he argued that Article 363 which excluded all covenants between the Government and the princes from the jurisdiction of the courts, should also be amended in light of the other proposed amendments. Desai argued that once the princes were 60  stripped of their titles they should have the fundamental right to go to the court like any other citizen of the country. The Swatantra Party and the Jana Sangh also opposed the Minister's proposed bill, while the two Communist Parties supported it by demanding that no compensation should be made to the princes at all. P. Ramamurthi of the Communist Party of India (Marxist) went so far as to claim that the President of the Republic should simply scrap all the purses and privileges by executive order under the terms of Clause 22 of Article 366 which indeed gave the President the power of from all taxes on income. Article 362. In the exercise of the power of Parliament or of the Legislature of a State to make laws or in the exercise of the executive power of the Union or of a State, due regard shall be had to the guarantee or assurance given under any such covenant or agreement as is referred to in Article 291 with respect to the personal rights, privileges or dignities of the Ruler of an Indian State. Article 366 (22). 'Ruler' in relation to an Indian State means the Prince, Chief or other person by whom any such covenant or agreement as is referred to in clause (1) of Article 291 was entered into and who for the time being is recognized by the President as the Ruler of the State, and includes any person who for the time being is recognized by the President as the successor of such Ruler, ^ h i s Article reads as follows: Article 363. Notwithstanding anything in the Constitution ...neither the Supreme Court nor any other court shall have jurisdiction in any dispute arising out of any provision of a treaty, agreement, covenant, engagement, sanad or other similar instrument which was entered into or executed before the commencement of the Constitution by any Ruler of an Indian State and to which the Government of the Dominion of India or any of its predecessor Governments was a party and which has or has been continued in operation after such commencement, or in any dispute in respect of any right accruing under or any liability or obligation arising out of any of the provisions of this Constitution relating to any such treaty ...  35  recognition. For a constitutional amendment bill to become law it had to be passed in both Houses of Parliament by a majority of the total membership and two-thirds of the members present and voting. On September 2,1970 the bill was put to a vote in the Lok Sabha. With 488 members of the total membership of 518 present, the government needed to secure at least 326 to have its bill passed. With a margin of just eight votes 61  the government managed to do so and the bill was passed in the Lok Sabha with 334 votes in favor and 154 against. The Congress (Ruling) was joined by the Dravida Munnetra Kazagham, the Communist Party of India, the Communist Party of India (Marxist), the Praja Socialist Party, the Samyukta Socialist Party, three members of the Bharatiya Kranti Dal, one member of the Swatantra Party and 26 Independents in supporting the bill. Those voting against the bill included the Congress (Opposition), the Swatantra Party, the Jana Sangh, the Akali Dal, eight members of the Bharatiya Kranti Dal, six members of the Congress (Ruling) and 14 Independents.  62  The passage of the bill in the 240 member Upper House, the Rajya Sabha, promised to be a much more difficult task. As a result of the Congress split, the governing party's share of the upper house seats dropped from 60% in August 1969 to 41.3% in February 1971, as shown in Table 2.3.  Two-thirds of 488 is 326 which is also greater than 259, which is one half of 518. 'Keesing's Contemporary Archives (1970), p. 24224.  36  T A B L E 2.3: THE PARTY-WISE COMPOSITION OF THE R A J Y A S A B H A , 1966-70 1969 August Party/Date 1966 1967 August 1968 June 1970 February INC/INC(R)  166 (69.7%)  INCO CPI(M)  166(69.2%)  140(58.3%)  144(60.0%)  99(41.3%)  4 (1.7%)  4(1.7%)  6(2.5%)  42(17.5%) 5(2.1%) 9 (3.8%)  CPI  8 (3.4%)  7(2.9%)  8 (3.3%)  9(3.8%)  PSP  6 (2.5%)  7 (2.9%)  4(1.7%)  4(1.7%)  BJS  6 (2.5%)  6 (2.5%)  11(4.6%)  10(4.2%)  10(4.2%)  3 (1.3%)  3 (1.3%)  3(1.3%)  BKD SWA  10 (4.2%)  13(5.4%)  16(6.7%)  13 (5.4%)  13 (5.4%)  DMK  2 (0.8%)  2 (0.8%)  3 (1.3%)  4(1.7%)  5(2.1%)  A KALI DAL  1 (0.4%)  0  1 (0.4%)  3(1.3%)  3 (1.3%)  INDEPS.  10 (4.2%)  7(2.9%)  9 (3.8%)  9(3.8%)  13 (5.4%)  OTHERS  13 (5.5%)  16(6.7%)  19(7.9%)  21 (8.8%)  24(10.0%)  NOMINATED  12 (5.0%)  11(4.6%)  12(5.0%)  12 (5.0%)  11(4.6%)  VACANCIES  4(1.7%)  1 (0.4%)  10 (4.2%)  2 (0.8%)  3 (1.3%)  TOTAL  238 (99.9%)  240 (100%)  240 (100.2%)  240 (100.3%)  240(100.3%)  Source: Butler at el., p. 63.  Moreover, the combined share of upper house seats held by those parties that supported the bill in the Lok Sabha was only 121 seats, or 50.4% which was sufficient to guarantee the passage of the bill in the upper chamber only if no more than 181 members turned out to vote. This situation provided an incentive for the opposition to ensure that the house 63  would be fully attended on the day of the vote: the more members the opposition could turn out over and above 181, the more difficult the government's position would be as it would have to engage in costly last-minute negotiations to secure the requisite support that the passage of the bill required. On the other hand, the governing party and its allies had to make sure both that all their members would be present and that they can persuade  Again, since the passage of the bill required the support of one-half of the total membership of the chamber and two-thirds of those present and voting, 121 votes would not have been sufficient if every member had turned out. In that case, a minimum of 161 votes would have been necessary. With 181 members voting, however, 121 votes, which is greater both than one half of 240 and than two-thirds of 181, would have been sufficient for the governing party and its supporters to have the bill passed. 63  37  as many opposition members as possible to abstain from attending.  64  The bill was put to a vote in the upper chamber on September 5, 1970 in the presence of 224 members. Thus, the bill needed to be supported by at least 150 members in order to pass. Although by and large political parties followed the same pattern of voting as they did in the Lok Sabha, there were minor deviations, which, in the end, made all the difference: each member of the Bharatiya Kranti Dal and one member of the Congress (Ruling) voted against the bill while the Akali Dal, two members of the Congress (Opposition) and one member of the Swatantra Party voted in favor of the bill. The tally was thus 149 in favor and 75 against the bill. Although the bill clearly enjoyed the support of the majority of the Rajya Sabha, by a margin of one vote it failed to be passed by two-thirds of the members present and voting. Following the failure to have the constitutional amendment bill passed, the Congress (Ruling) cabinet decided to hold an emergency meeting. It was decided that notwithstanding the failure of the bill, the President should de-recognize princely privileges by an executive order under the terms of Clause 22 of Article 366. The same evening the President's signature was obtained and de-recognition took effect. Justifying this measure, Y . B. Chavan told the Rajya Sabha: The President has under the Constitution the unquestioned power to derecognize the rulers. During the debate in parliament it was even urged by some members [read the Communists] that the Government should have resorted to this power instead of bringing in the Constitution Amendment Bill.... The Government is fortified in the belief that there is widespread support in the country for putting an end to an outmoded and antiquated system which permitted the enjoyment of privileges and privy purses by a small section of our people without any corresponding social obligation on their part.... During the course of the debate in Parliament even many of those who opposed the Constitution Amendment Bill expressed themselves in favor of the abolition of privileges and privy purses. 65  In its frantic effort to ensure full attendance the Jana Sangh had one of its members brought into the chamber from the hospital on a stretcher, while a member of the Samyukta Socialist Party was escorted by a police guard due to having been recently arrested. Keesing's Contemporary Archives (1970), p. 24224. 64  65  38  'Almost immediately, seven of the now de-recognized rulers submitted a petition to the Supreme Court against the President's order. In its ruling on December 15, 1970 66  the Court rendered the President's order ultra vires the Constitution and restored the rights and privileges of the princes. In two weeks' time, on December 27, upon the 67  advice of Prime Minister Indira Gandhi, the President dissolved the Lok Sabha and ordered fresh elections although they were not due to be held until early 1972. In statements after the announcement was made, Indira Gandhi acknowledged that her party wanted to end the situation of having to rely on the support of small parties in order to provide effective governance. A new election, to be fought on the issue of social justice and egalitarianism of which the Congress (Ruling) projected itself to be the champion, might give the party the opportunity to regain its majority.  68  A n important consequence of the early dissolution of the Lok Sabha was the decoupling of national and subnational elections. In the past, elections for both the Lok Sabha and the Vidhan Sabhas had been held concurrently and therefore, their terms would also expire at the same time. From time to time, there would be a few states where elections would have to be held off-schedule because of the imposition of President's Rule or the dissolution of the state assembly by the central government. However, as Table 2.4 shows, until 1967 most Vidhan Sabhas had been elected simultaneously with the Lok Sabha.  'Keesing's Contemporary Archives (1970), p. 24224..  Tbid, p. 24409. 'Ibid., p. 24427.  39  T A B L E 2.4: CONCURRENT A N D NON-CONCURRENT NATIONAL A N D SUBNATIONAL ELECTIONS IN INDIA, 1952-1996 # of subnational # of subnational # of subnational legislatures units with units with nonLok Sabha election concurrent Vidhan concurrent Vidhan year Sabha election Sabha election 1952 13 13 0 1957 13 13 0 1962 16 14 2 1967 23 21 2 1971 23 3 20 1977 25 1 24 1980 27 6 21 1984 27 5 22 1989 27 5 22 ' 1991 27 7 20 1996 27 6 21 Notes: Subnational unit refers to both states, territorial councils ,and union territories that had popularly elected legislatures at the time. In 1962 there were no concurrent elections in Kerala and Orissa, in 1967 in Nagaland and Pondicherry. The states where there were concurrent elections after 1971 are: 1971: Orissa, Tamil Nadu and West Bengal; 1977: Kerala; 1980: Manipur, Arunachal Pradesh, Goa-Daman and Diu, and Pondicherry; 1984: Tamil Nadu, Manipur, Arunachal Pradesh, Goa-Daman and Diu; 1989: Andhra Pradesh, Goa, Karnataka, Sikkim, and Uttar Pradesh; 1991: Assam, West Bengal, Uttar Pradesh, Haryana, Kerala, Tamil Nadu, and Pondicherry; 1996: Assam, West Bengal, Haryana, Kerala, Tamil Nadu, and Pondicherry. Sources: Butler et al., India Decides... various tables, and Aggarwal and Chowdhry, Elections in India pp.52-59.  2.3  T H E FIFTH L O K S A B H A , 1971-7: E M E R G E N C Y R U L E A N D ITS CONSEQUENCES The fifth general election vindicated Indira Gandhi as she led the  Congress(Ruling) to a massive victory. The party won 44% of the popular vote and 68% of the Lok Sabha seats. It was indicative of the party's strength that it could stage this remarkable performance without engaging in electoral coordination with other political parties on a national scale. Locally, however, the Congress(Ruling) did make seat adjustments with various parties. For example, in the state of Kerala, the Congress (Ruling) and its partner in the United Front coalition cabinet, the Kerala Congress,  40  entered into an agreement not to field candidates against one another. Mrs. Gandhi's party concluded similar deals with the Dravida Munnetra Kazagham in Tamil Nadu, and with the Communist Party of India in a number of different states.  69  Of the opposition parties, the Opposition Congress, the Samyukta Socialist Party and the Bharatiya Jana Sangh formed a national alliance that would field one jointly supported candidate in each constituency. The Swatantra Party was hesitant about joining, however, after having been assured that the alliance would not form an electoral pact of any sort with the Communist Parties, it decided to cooperate. This caused some uneasiness in the alliance because Madhu Limaye, leader of the Samyukta Socialist Party, claimed that his party had entered the alliance in the first place on the basis of the understanding that it would be allowed to engage in electoral coordination with the Communist parties on its own.  70  With the solid majority of the Congress(Ruling), the Fifth Lok Sabha promised governmental stability. Ironically, however, it was exactly during the term of this legislature that the suspension of democracy and the imposition of Emergency rule, the most dramatic and traumatic event in the history of India's party politics, took place. The new Congress(Ruling) administration found itself unable to cope effectively with rising inflation and increasing food shortages, both caused by the war with Bangladesh, the drought of 1972-73 and the world energy crisis. The economic crisis 71  quickly led to a political one as anti-government protests and demonstrations were becoming an every-day occurrence all over the country. Political instability was at its worst in the states of Gujarat and Bihar. In the former, a massive student agitation led to the resignation of the state's Congress(Ruling) government and the imposition of President's Rule on the state. In Bihar, a movement led by Jayaprakash Narayan went  Keesing's Contemporary Archives (1971), p. 24581. Ibid Ffardgrave and Kochanek, India: Government and Politics,  69 n  71  p. 259.  41  beyond demanding the resignation of the state Congress(Ruling) government and called for a total transformation of the entire society, a total revolution that would benefit the disadvantaged sectors of the society.  72  The crisis that beset the national government was further compounded by the delivery of the Allahabad Election Tribunal's verdict on June 12, 1975 in the Raj Narain versus Indira Gandhi case. Mr. Raj Narain, who had been defeated by Prime Minister Indira Gandhi in the Rae Bareli electoral district in Uttar Pradesh state in the 1971 general election, sued the Prime Minister for corrupt electoral practices. The tribunal found her guilty and consequently ruled to unseat and disqualify her from participating in electoral activity for six years. The Prime Minister's subsequent application to the Supreme Court of India for an unconditional stay of the judgment was turned down on June 24 by Justice Krishna Iyer, who, however, also ruled that the Prime Minister could retain her seat but could neither vote nor participate in parliamentary proceedings until the Court would properly convene and rule in her case.  73  Justice Iyer's decision was announced on the same day when the Gujarat state elections were concluded. This state election was won by the Janata Morcha, an electoral alliance formed by the Congress (Organization), the Jana Sangh, the Bharatiya Lok Dal, the Socialists, and the National Labor Party (Majur Mahajan) who now felt encouraged to work towards forming a similar alliance for the next national election which was less than a year away. The following day, on June 25, a massive rally was organized in New 74  Delhi under the leadership of Morarji Desai and Jayaprakash Narayan to demand the immediate resignation of the Prime Minister. In response, upon Indira Gandhi's advice, the President of the Republic, Fakhruddin Ali Ahmed issued a proclamation of Ghansayam Shah, Protest Movements in Two Indian States: A Study of the Gujarat and Bihar Movements (Delhi: Ajanta Publication, 1977). F£ardgrave and Kochanek, India: Government and Politics, p.260. Although the Janata Morcha did not win a majority of the seats in the Gujarat Assembly, it was able to form a government with the assistance of the Kisan Mazdoor Lok Paksha led by Chimanbhai Patel. 72  73  74  42  Emergency on June 26, 1975. The President convened an Emergency session of the lower house where the Speaker suspended regular parliamentary proceedings and ruled that only government business would be transacted during this session. This made it very easy for the Congress government to have three constitutional amendment bills, each aiming at increasing the autonomy of the executivefromboth the legislature and the judiciary, passed. In particular, the 39th Constitutional Amendment made the President's proclamation of Emergency non-challengable in any court of law. In addition, it also stipulated that the President may issue multiple proclamations at the same time. The 40th Amendment sought the creation of a parliamentary authority with exclusive jurisdiction to try matters pertaining to the election of the President, the Vice President, the Prime Minister and the Speaker of the lower house. The Amendment made the law establishing this authority non-challengeable in any court of law. The 41st Amendment extended the immunity the President had enjoyed from criminal investigations to covering the Prime Minister. Besides these amendment bills, the government also secured the passage of a number of laws that enabled it to incarcerate opposition leaders and enhance its control over the polity. Among the opposition parties only the CPI hailed the imposition of the Emergency. The CPI was keen on retaining the support of the Congress Party for the sake of continuing their coalition in the state of Kerala where the CPI supplied the head of the state government in spite of its junior status. The CPI(M), which had constituted the united Communist Party of India with the CPI until their split in 1964, opposed the Emergency. However, the party did not join the BLD, the Socialist Party, the BJS and the Congress(O) when they agreed on January 4, 1976, the day before the commencement of the Winter session of parliament, to form a joint parliamentary bloc, called the Janata Front, to coordinate their legislative efforts against the Congress(Ruling) in both houses of parliament. This move towards creating a jointfrontagainst the government was  43  especially important because there were no recognized opposition parties in the lower house which gave the governing party a major procedural and tactical advantage during house deliberations." On January 5, 1976 the cabinet resolved to advise the President to extend the term of the lower house by one year. This meant that elections would not be held as 76  scheduled in March 1976 when the regular term of the house expired. The government dealt a similar blow to the Tamil Nadu and Gujarat State Assemblies, the only two states where the Congress(Ruling) was not in office. Although the term of the Tamil Nadu Legislative Assembly was due to expire in March 1976 concurrently with the original term of the Lok Sabha, the Congress (Ruling) government advised the President to dissolve it and impose President's Rule on the basis of corruption charges that the state governor filed about the state government. In Gujarat, in early March 1976 the KMLP, on whose external support the Janata Morcha coalition depended, dissolved itself. With most of the 12 KMLP legislators joining the Congress(Ruling) Party and with a number of Janata Morcha legislators crossing the floor, the government lost a budgetary vote and resigned. President's Rule was proclaimed on March 12 although the state assembly was only suspended but not dissolved. According to house rules, a parliamentary group is accorded recognition as a party, and consequently its leader becomes the leader of the Opposition, by meeting three criteria. First, it must comprise at least one tenth of the membership of the house so that it could constitute quorum on its own; second, it must subscribe to a set of clearly identified ideological principles; and third, it must have an extra-parliamentary organization. Being a recognized parliamentary party as opposed to merely being a group has two main advantages. First, only recognized parties are being allotted office space in the house which makes their presence and liaison permanent even between parliamentary sessions. Second, the leader of the opposition has some influence over the legislative agenda in terms of tabling motions, consulting with the Speaker about the order of the day etc. However, as long as there is no recognized opposition party in the house, the input of the opposition in parliamentary proceedings can be much more easily manipulated by the Speaker. "According to Article 83 of the Constitution, the President could extend the term of the lower house under Emergency for one year at a time but it cannot be extended beyond six months after the proclamation has been lifted. 75  44  In the wake of these events, the BLD, the Socialist Party, the Jana Sangh and Congress(Organization) held a meeting in New Delhi where they adopted a statement that went beyond the idea of a unitedfrontby calling for the creation of a united party whose primary objective should be "the restoration of democracy, democratic values and civil liberties" through peaceful means . The meeting set up a Steering Committee comprising 77  one memberfromeach of the four parties and instructed it to complete the integration of the four parties in two months. During these two months, the respective decision-making authorities of the four parties deliberated on the merger proposal. The Working Committee of the Congress(O) arrived at a consensus that "the Congress-0 should neither merge with the Ruling Congress nor with any other party, nor should it dissolve itself to form a united opposition party." At the same time, the Committee also recognized that coordinating candidate nomination and electoral strategy with the other parties was very important in order to avoid the splitting of the anti-Congress vote. Similarly, the Working Committee 78  of the BJS passed a resolution stating that the party was generally in favor of the merger but that itsfinaldecision in this regard should wait until all the party's leaders were releasedfromjail. The Socialist Party insisted that before a complete merging of the parties' identities took place their respective auxiliary organizations must also be satisfactorily reconciled. Of the four parties, the BLD was the only one that consistently maintained a position in favor of the merger. In doing so, however, the party 79  emphasized that it wanted the new party to adopt the BLD's industrial and agricultural Madhu Limaye, Janata Party Experiment. An Insider's Account of Opposition Politics: (Delhi: B.R. Publishing Corporation, 1994) pp. 125-6. Limaye, p. 131. The BLD itself came into existence in May 1974 as a result of the merger of seven parties. These parties were the Bharatiya Kranti Dal, the Swatantra Party, a faction of the Samyukta Socialist Party led by Raj Narain and Karpoori Thakur, the Utkal Congress, the Kisan Mazdoor Party, the Kshatriya Loktantrik Sangh, and the Punjab Farmers' Union. In September 1974, a faction of the United Goans Party also joined the BLD. See, Political Parties of Asia and the Pacific (Westport: Greenwood press, 1985), p.304. 77  1975-1977  78  79  45  policy. Frustrated with the indeterminate outcome of the series o f meetings among the four parties, the Congress(O) and the B L D engaged in bilateral talks about the merger during the fall o f 1976. The Congress(O) wanted the new party to be organized according to its own constitutional structure and retain the word Congress in the new party's name. The B L D in exchange wanted to ensure that its agricultural and industrial policies would be accorded priority among the new party's objectives and also wanted to fill the new party's presidency. However, the agreement failed to be realized because the Congress(O) Working Committee was similarly adamant about providing the new party's president. O n October 30, 1976 the President, on the Prime Minister's advice, ordered another one-year extension of the term of the lower house. However, in the hope o f catching the opposition unprepared, the Prime Minister suddenly advised the President to dissolve the L o k Sabha on 18 January 1977 and order new elections to commence on March 16, 1977. The Janata Front applied to the Election Commission for a new symbol 80  on which to contest the election, however, its application was turned down. 80  81  A t the same  In the meanwhile, the Congress(Ruling) government secured the passage o f the 42nd  Constitutional Amendment Bill in the L o k Sabha on November 11. The Bill intended to further strengthen the executive vis-a-vis both the opposition and the other branches o f government. Its most important provisions were the following: a ban on anti-nationalist organizations; the President was bound to act on the advice of the Council o f Ministers; the regular term o f the L o k Sabha and the Vidhan Sabhas was increased from five to six years; the President may disqualify members o f parliament and members o f legislative assemblies on the grounds of corruption changes; only the Supreme Court may rule a law enacted by the central government to be unconstitutional and in order to do there has to be a 2/3 majority on the Court; the central government may send armed forces into the states without the consent o f the state government; the President may proclaim Emergency over selected parts of the country; no constitutional amendment was subject to judicial review; President's Rule may be imposed on a state legislature for one year without the approval of the national parliament; the President may repeal a constitutional amendment within two years of its enactment by proclamation. 81  Party symbols play a very important role in elections because o f the high rate o f  illiteracy among the electorate. Symbols are allotted to recognized political parties only. Recognition o f political parties was regulated by the Election Symbols (Reservation and  46  time, the Commission maintained that nothing prevented the Front members from running their candidates on a symbol already owned by a recognized party. Without recognition, however, the candidates run by the Front could only be classified as Independents. In the light of the Commission's ruling, the four parties agreed that their candidates would contest the election on the BLD symbol all over the country with the exception of the southern state of Tamil Nadu where the Congress(O) symbol would be used. They also established a 27-member National Executive to take charge of the affairs Allotment) Order of 1968. According to paragraph 2 of the Order, a political party shall be treated as a recognized political party in a State, if and only if either the conditions specified in clause (A) are, or the condition specified in clause (B) is fulfilled by that party and not otherwise, that is to say — (A)  that such party — (a) has been engaged in political activity for a continuous period of five  years; and (b) has, at the general election in that State to the House of the people, or, as the case may be, to the Legislative Assembly, for the time being in existence and functioning returned — either (i) at least one member to the House of people for every twentyfive members of that House or any fraction of that number electedfromthat State; or (ii) at least one member to the legislative Assembly of that State for every thirty members of that Assembly or any fraction of that number; (B) that the total number of valid votes by all the contesting candidates set up by such party at the general election in the State to the House of the people, or as the case may be, to the Legislative Assembly for the time being in existence and functioning (excluding the valid votes of each such contesting candidate in a constituency as has not been elected and has not polled at least one-twelfth of the total number of valid votes polled by all the contesting candidates in that constituency) is not less than four per cent of the total number of valid votes polled by all the contesting candidates at such general election in the State (including the valid votes of those contesting candidates who have forfeited their deposits.) The Order further points out that if a party gains recognition according to these rules in four or more states then it is classified a national party and can use the same symbol throughout the entire county. Other recognized parties can use their symbol only in the states where they are recognized.  47  of the new party which was agreed to be officially inaugurated under the name Janata Party after the election. On February 2, Defense Minister Jagjivan Ram resignedfromthe CongressfRuling) cabinet and formed a new party, the Congress For Democracy (CFD). The CFD entered into electoral coordination with the Janata and alsofieldedits candidates under the BLD symbol. 2.4  THE SIXTH LOK SABHA, 1977-80: TWO UNSTABLE COALITION GOVERNMENTS The general election resulted in a narrow Janata majority as Table 2.5 shows.  Although the five original constituents of the Janata Party had contested the elections on the same symbol, except in the state of Tamil Nadu, according to unofficial reports the approximate seat shares won by the Janata's respective components were as follows: the Jana Sangh had about 90 seats, the Bharatiya Lok Dal had 68, the Congress(O) had 55, the Socialists had 51, the Congress for Democracy had 28 and the remaining 6 seats of the Party were held by dissident individual ex-Congressmen.  82  TABLE 2.5: THE SIXTH LOK SABHA, 1977-80, BEFORE AND AFTER THE JANATA PARTY # of seats after % of seats after # of seats after Party the election the election Janata split Janata Party 295 54.60% 206 Janata Party (S) N A NA 76 Congress 154 28.50% 75 Congress(T) NA NA 71 AIADMK 19 3.50% 19 CPI(M) 22 4.10% 22 SAD 9 1.70% 9 Others 32 5.90% 27 Independents 9 1.70% 33 Total 540 100% 538  The Hindu (June 24, 1978)  SPLIT OF THE % of seats after Janata split 38.30% 14.10% 13.90% 13.20% 3.50% 4.10% 1.70% 5.00% 6.10% 99.90%  48  On March 24 the Janata parliamentary group selected Morarji Desai, the leader of the Congress(O), as its Prime Ministerial nominee who was then duly invited by the President to form a government by virtue of leading the largest party in the house. The 83  four-party alliance and the CFD agreed that the new cabinet would comprise 2 members from each of the five parties as well asfromthe Shiromani Akali Dal that contested the election in alliance with the Janata in the state of Punjab. However, the new Prime Minister submitted a list of 19 cabinet members to the President. The list gave 6 portfolios to the Congress(O), 4 to the BLD, 3 each to the BJS and the Socialists, 2 to the CFD and 1 to the SAD. In the aftermath of the election, the CFD tried to win over legislatorsfromthe defeated Congress(Ruling) Party in order to increase its weight in the Janata coalition which was particularly in need of improving its position in the upper house where the CongressfRuling) Party still had an overwhelming majority, as Table 2.6 shows. However, the Prime Minister stood resolutely against admitting Congress(Ruling) defectors in the coalition. TABLE 2.6: THE PARTY-WISE COMPOSITION OF THE RAJYA SABHA AS OF MARCH 1976  PARTY Congress Party CongressCO") CPT CPIuVD PSP BJS BKD n w Akali D a l  Independents Others Nominated Vacancies: Total  NUMBER OF SEATS 142 7 1?. 5 1 12 11 Q  2 15 8  12  PERCENT OF SEATS 58.20% 2.90% 4 90% 2.00% 0.40% 4.90% 4.50% 7 70% i ?n%  080% 6.10% 7 7fW„  4.90% ?nn% inn%  Source: Butler et al., India Decides..., p. 63  'Officially, the Janata Party did not come into existence until only May 1, 1977.  49  Although the entire tenure of the Janata Party in office was characterized by an uneasy relationship among its original constituents, the process leading up to the cabinet's collapse started with Prime Minister Morarji Desai's decision to remove Charan Singh, leader of the BLD faction,fromthe Home Ministry and expel himfromthe cabinet in June 1978. According to Desai, the reason for this decision lay in Singh's criticism of the cabinet in public which broke the principle of collective responsibility. According to Singh, however, the real reason was twofold:first,as Home Minister he had led an investigation into the alleged corrupt practices of the Prime Minister's son, Kanti Desai. Second, Singh believed that he was also becoming too much of a threat to the Prime Minister's position. To mark his 77th birthday, Singh and his supporters organized a mass peasant rally in the heart of New Delhi on December 23, 1977 as a show of strength. The rally was supported by the erstwhile Bharatiya Kranti Dal and the Akali Dal, the Janata Party's coalition partner, whose leader Prakash Singh Badal was holding the Agriculture portfolio. On January 24 1979 Singh was re-admitted into the cabinet as Deputy Premier and Finance Minister. The Prime Minister also appointed Jagjivan Ram, the leader of the Congress for Democracy, as Deputy Premier and left him in charge of the Defense Ministry. In December 1978, O. P. Tyagi, a member of the Jana Sangh faction, introduced the Freedom of Religion Bill, in the Lok Sabha. The bill which sought to make religious conversions by the use of "force, fraud and inducement" a criminal offense outraged the Christian community which played a significant role in running and providing for schools and hospitals, which could be easily construed to be inducements. Although the Prime Minister dropped the bill in May 1979 after a mass demonstration by a Christian group in New Delhi, he claimed that his government would soon re-introduce a similar bill in the  50 future.  84  In April, the Prime Minister made yet another religion-sensitive announcement when he proposed to introduce a constitutional amendment bill in order to make cow protection a concurrent subject, i.e. an issue area about which both the state and the national governments could legislate. Furthermore, he announced that this amendment would then soon be followed by a law that would ban cow slaughter in the entire country.  85  This proposal was sensitive because for Hindus, constituting the religious majority in the country, the slaughtering of cows is sacrilege, however, cow eating was perfectly acceptable for the other religious communities. Thus, by banning the killing o f cows the government was in effect proposing to impose Hindu dietary habits on the minority communities which was in violation of the country's constitutionally guaranteed secular principles. Furthermore, as critics pointed it out, cow protection made no economic sense as it already resulted in millions of old and diseased cows wasting land and fodder and reducing the milk-yield of the country's live cow stock. In addition, for many poor people, cow meat was the only affordable high-protein food stuff. A t the time this Constitutional Amendment Bill was proposed, in only two states, Kerala and West Bengal was cow slaughter permitted. It is important to note that it was in these two states that the two Communist parties, the arch enemies o f the Jana Sangh, were very strong. O n M a y 18, 1979 the L o k Sabha passed the bill despite the opposition o f the Communist parties, the Congress, the Muslim League, the A D M K , the Forward Bloc and even the Socialist faction of the Janata Party itself.  86  The second half o f 1978 and the first half o f 1979 were marked by a series o f some o f the worst communal riots in a long time between Hindus and Muslims. The  Keesing's Contemporary Archives (1979), p. 29969. Keesing's Contemporary Archives (1979), p. 29969.  M Si  mid  51  incident led to repeated demands by members of the BLD and the Socialist factions of the Janata Party that the Janata should sever all ties with the paramilitary Hindu fundamentalist organization, the Rashtriya Swayamsewak Sangh (RSS). This, however, was easier said than done because of the Jana Sangh's close relationship with that organization.  87  As a direct outcome of the tension between the Jana Sangh and the rest of the Janata factions, the governments of the states of Uttar Pradesh, Bihar and Haryana, each of which w a s dominated by the Bharatiya Lok Dal group, collapsed. In Uttar Pradesh, Chief Minister N. R. Yadav dismissed all 15 Jana Sangh ministersfromhis cabinet. In a subsequent vote of confidence taken by the Janata Legislature Party, he lost and resigned. His successor, Banarasi Das, was a member of the Congress(O) faction who appointed a cabinet including no Jana Sangh member. In the state of Bihar, Chief Minister Karpoori Thakur resigned after the Jana Sangh engineered his defeat on a confidence vote held by the Janata Legislature Party. The next Chief Minister, Ram Sunder Das, belonged to the Socialist faction. His cabinet included members from each of the Janata constituents with the sole exception of Thakur's BLD. Finally, in Haryana, Chief Minister Devi Lai was forced to seek a vote of confidence in the Legislature Party after he dismissed his cabinet's Jana Sangh ministers. Instead of seeking a vote, Lai resigned and was succeeded by Bhajan Lai, a member of the Congress for Democracy group supported by the Jana Sangh.  88  Following these changes in the state governments, the BLD increased its efforts against the Jana Sangh in the national party. On April 2, Raj Narain, a senior BLD leader, demanded once again that the Party officially break all ties with the RSS. In protest, the Jana Sangh group retorted that should the Party proceed to do so, the Jana Sangh would  Being a very well organized paramilitary group the RSS provided indispensable assistance to the BJS in terms of electoral mobilization. **Keesing's Contemporary Archives (1979), p. 29970. 87  52  exitfromthe Janata Party. Furthermore, the Jana Sangh gathered the support of 100 Janata Party MPs demanding the expulsion of Raj Narain and Banarasi Das, the BLD's only remaining Chief Ministers,fromthe party organization.  89  On May 17, a conference was convened by Madhu Limaye, a member of the Socialist faction and one of the general secretaries of the Janata Party, about the possibility of forming a "third force" that would be opposed to both the authoritarianism of Indira Gandhi and the communalism of the RSS. The conference was attended by representatives of the Congress, the two Communist Parties, the Forward Bloc, the Revolutionary Socialist Party, and the Peasant's and Workers' Party and the anti-RSS members of the Janata Party. These parties resolved to commit themselves to cooperating in the Lok Sabha "on issues of common concern such as national unity, democratic freedoms, communal harmony and social justice". Soon after the conference, Raj Narain 90  left the Janata Party and started a new parliamentary group called the Janata Party (Secular). On July 10, Y. B. Chavan, the Leader of the Opposition, tabled a no-confidence motion against the government accusing it of tearing apart the country's secular fabric. Meanwhile, an increasing number of BLD and Socialist members of the Janata defected to join Raj Narain's' group in the House. By July 11 the number of defectors reached 49, depriving the Janata cabinet of its majority in the Lok Sabha. At this point, for the 91  government to survive it had to have the support of both the Anna-DMK and the 92  Communist Party of India (Marxist), see Table 2.5. Although the former did lend its support to the Janata Party, the latter resolved that it would never support a party in government that contained communal elements in its ranks.  "Ibid, p. 29971. ^Keesing's Contemporary Archives ( 1979), p. 29971.  "Ibid. ^The Anna-DMK was formed by defectors from the DMK led by party treasurer M. G. Ramachandran in 1972.  53  In the evening of July 15, 1979 the day before the vote on the no-confidence motion was due to be held, Prime Minister Morarji Desai submitted his resignation to the President Sanjiva Reddy who asked him to remain in office in a caretaker capacity until a new government was formed. By this time, the Congress(Ruling) had also split as a result of the deepening rift between those Congressites loyal to former Prime Minister Indira Gandhi and those who blamed the party's rout in the 1977 elections on her Emergency regime. The latter group was led by Brahmanand Reddy, and although it was in a minority in the All-India Congress Committee, it did control a majority of the Congress(Ruling)'s members of parliament. In contrast, Indira Gandhi's group was stronger in the party organization. The Reddy-group, led by Y. B. Chavan in the Lok 93  Sabha, expelled Indira Gandhi from the party, in reaction to which, on January 2, 1978, Indira Gandhi and her loyal supporters convened to form the Congress(I).  94  Following the resignation of Morarji Desai, President Reddy invited Chavan, as the leader of the largest opposition party, to form a government on July 18. As a result of the various permutations that had taken place within and among the different parties the composition of the Lok Sabha changed significantly since the election as shown in Table 2.5. The Janata Party now had only 206 seats in contrast with the 298 it had won in 1977, the Janata Party (Secular) had 76 seats, the Congress 75, the Congress(I) 71, and 11 smaller parties accounted for 77 seats while there were 33 Independents as opposed to only 9 that had been elected as such. Since five seats were vacant, for a new government to have a majority it would have had to control at least 270 seats. It is worth mentioning that Indira Gandhi herself was expelled from the Lok Sabha after the House Privileges Committee ruled that she was guilty of breach of privilege and contempt of the Lok Sabha as she had misused her Prime Ministerial office to protect her son's business dealings. The Lok Sabha accepted the report and passed a motion proposed by Prime Minister Morarji Desai to expel Indira Gandhifromthe House on December 19, 1978. In a later by-election Indira Gandhi regained her seat by winning the Chikmagalur constituency in Karnataka. See, Keesing's Contemporary Archives (1979), p. 29965. The "F in the brackets stood for nothing other than the leader of the new party, Indira! 93  M  54  The same day, the Janata Parliamentary Party executive held a meeting where suggestions had been made to remove Morarji Desai from the party's leadership.  95  Jayaprakash Narayan himself, the revered leader of the movement that led to the creation of the Janata Party, extended an appeal to Desai to step down and give way to Jagjivan Ram. However, Desai refused to resign and stayed on as the leader of the Janata Parliamentary Party. On July 22, Chavan advised the President that he was unable to form a government. However, he had formed an alliance with Charan Singh's Janata Party (Secular) with the objective of forming a coalition government in which Charan Singh would be the Prime Minister and Chavan the Deputy Prime Minister. In the meanwhile, Morarji Desai also staked a claim to form a government on behalf of the Janata Party. In turn, the President asked both Desai and Singh to provide him with a list of supporters within two days in order to determine who would be in a position to form a majoritybased government. During those two days, the Janata Party (Secular)-Congress(Reddy) alliance was promised the support of the Congress(I), a 15-member Socialist group, the CPI, the Muslim League and the Peasants' and Workers' Party. The Janata Party was promised to be supported by only the ADMK. The CPI(M), the Forward Bloc and the Revolutionary Socialist party remained non-committal.  96  The support that the Congress(I) extended to the Janata Party (Secular)-Congress alliance was received uneasily by the Socialists who had declared on July 24 that they would not support a government that depended on the assistance of the Congress(I). In an effort to deflect the Socialists' criticism, Charan Singh issued a communique clarifying his position concerning the acceptance of the Congress(I)'s assistance: In this task of forming a Government, I have relied mainly on my colleagues of the Janata (S) and the Congress as well as the progressive forces. I would like to make it clear that the support of Cong.(I) came to me unconditionally while I and my colleagues Madhu Limaye, The Great Janata Experiment, pp. 497-99. ^Keesing's Contemporary Archives (1979), p. 29972. 95  55  were engaged in the task of forming a viable, progressive, democratic and secular government. It would indeed be unfortunate if in this great task of giving the nation a truly dynamic Government, support of Cong. (I) from outside is made much of. This may lead to reinstalling a Government which would be totally under the baleful influence of fascist communal elements.  I would like to assure that nothing has been done by me to compro relentlessfightagainst communalism and authoritarianism. I, therefore, appeal to all my colleagues in the parliament to consider the situation with great care. 97  On July 24^Singh and Desai submitted their respective lists of supporters to the President. However, both lists contained the names of 279 supporters with 42 being included in both. Curiously, both lists included the Akali Dal and Desai's list also included 21 Congressmen who later denied that they had given their consent to be listed as Janata supporters. After closer inspection, the President determined that the Janata Party had the legitimate support of only 238 MPs while the Janata Party (SecularCongress combine was supported by 262 MPs. On July 26, President Reddy invited Charan Singh to form a government even though he could not prove with certainty that he would have majority support in the Lok Sabha. The President instructed Singh to seek a vote of confidence by the third week of August. On July 28, Charan Singh was sworn in as Prime Minister with his cabinet comprising six ministers from the Janata Party (Secular), seven from the Congress, one from the Socialist group and onefromthe CFDS. Thus, a coalition minority government was formed.  98  As soon as the cabinet was sworn in, the coalition partners engaged in a frantic search for allies and additional participants in order to survive the confidence vote. On August 8, 1979 the ADMK announced that it would support the government and eleven days later two of its members were inducted in the cabinet: Aravinda Bala Pajanor received the Petroleum and Chemicals portfolio and Sathiyavani Muthu was given the ministry of Social Welfare. The Communist Party of India, the Communist Party of India Limaye, The Great Janata Experiment, pp. 509-10. ^Keesing's Contemporary Archives, p. 29972. 97  56  (Marxist) as well as a number of small leftist parties also pledged their support to the coalition on August 16." However, there were two developments that were gradually overshadowing the apparent progress that the coalition was making in terms of expanding and consolidating its majority in the House. The first was the split in the CongressfReddy). On August 19, just one day before the vote of confidence was scheduled to be taken, eleven members of the Congress(Reddy) Parliamentary Party resigned protesting against the underrepresentation of social minorities among the Congress ministers who were given portfolios in the coalition cabinet. Although the Congress ministers submitted their resignations in response to these protests, the dissidents broke away from the Party and formed the Indian National Congress (Real) threatening to vote against the government the next day.  100  A more troubling development was the changing attitude of the Congress(I) towards the coalition. On July 29, the day after the cabinet was swom in, the Congress(I) announced that it had provided unconditional support only for the formation of the coalition cabinet, however, as far as the continuation of the government in office was concerned the Congress(I)'s support "was not totally unconditional." In particular, the 101  Congressfl) specified two conditions for the continuation of its support. First, the coalition would have to consult the Congress(I) on every major policy decision, and second, no Congress member might be included in the cabinet who had been an opponent of Indira Gandhi in the past. The latter condition was a veiled demand to remove two 102  cabinet ministers in particular from their posts, T. A. Pai and Karan Singh, both of whom had made depositions against Indira Gandhi before the Shah commission that had been  "Sharda Paul, 1980 General Elections in India (New Delhi, Associated Publishing House, 1980), p. 67. Also see, Keesing's Contemporary Archives (1979), p. 29973. ^Keesing's Contemporary Archives, p. 29973. Paul, 1980 General Elections in India, p.66. ^Keesing's' Contemporary Archives, p. 29973. 101  57  appointed by the Desai government to inquire into excesses committed during the Emergency. Although it was not until the end of July that the Congress(I) would start criticizing the government openly, the party's commitment to supporting the coalition was never really credible. Thefirstindication of this was revealed by President Reddy who had informed Madhu Limaye, one of the general secretaries of the Janata Party and leader of the Socialist faction, that no sooner had the crisis caused by the split of the Janata Party broken out than Indira Gandhi indicated to him that her party was aiming at a mid-term poll and that the Congress(I)'s support to the Janata Party(Secular)Congress(Reddy) coalition was merely a stop-gap measure.  103  In addition, while extending support to the coalition at the national level, the Congress(I) cooperated with the Janata Party against the Janata Party(Secular) in the Bihar and Uttar Pradesh State Assemblies. In the Bihar Vidhan Sabha, the Janata Party (Secular) Legislature Party, led by former Chief Minister Karpoori Thakur, had submitted a motion of no-confidence against Chief Minister Ram Sunder Das' Jana Sanghdominated Janata cabinet. Although Thakur expected the Congress(I)'s cooperation so that they would jointly oust the governmentfromoffice, the latter refused to do so and voted against the motion, saving the Janata government. Similarly, in the Uttar Pradesh Assembly, the Congress(I) also refused to cooperate with the Janata Party (Secular). The Party joined hands with the Jana Sangh-led Janata Legislature Party in an effort to topple the Janata Party (Secular) government of the state headed by Chief Minister Banarasi Das. In this case, however, the Congress(I) and the Janata failed in their effort as the Das government survived the censure motion by a narrow margin of 21 votes.  104  The same day when the censure vote was taken in the Bihar Assembly, C. M. Stephen, the leader of the Congress(I) Parliamentary Party, met Jagjivan Ram, the new Limaye, The Collapse of the Janata Experiment ...p. 513. Ibid.,p. 515.  103  m  58  leader of the Janata Party who had recently replaced ex-Prime Minister Morarji Desai. The two leaders agreed that their parties would coordinate their legislative strategies in the Lok Sabha as well. Stephen once again confirmed that the Congress(I)'s support to the coalition pertained only to the formation but not necessarily to the continuation of the cabinet in office. On August 6, Jagjivan Ram announced that his party would welcome the Congress(I)'s assistance in toppling the Janata Party (Secular)-Congress coalition cabinet. He further stated that 105  When this government falls, the President will have no option but to invite the Leader of the Opposition [read Jagjivan Ram] to explore the possibility of forming a new Government, and I am confident that by then we would have sufficient strength to form a stable government. Let those Congressmen who have revolted come together, and I invite them to come under the banner of my party.  106  In the morning of August 20, the day of the confidence vote, the Congress(I) announced that it would vote against the coalition. Since the Congress(I) and the Janata Party held a majority of the seats in the House, it was certain that the coalition cabinet would not survive the vote. Prime Minister Charan Singh submitted his resignation to the President and advised him to dissolve the Lok Sabha and order mid-term elections. Jagjivan Ram called on the President on August 22 staking a claim to form a government. Ram indicated that since Charan Singh had never enjoyed the confidence of the Lok Sabha he was not a Prime Minister proper and therefore his advice was not binding on the President, i.e. the House ought not to be dissolved. Furthermore, Ram also indicated to the President that with the support of the ADMK and other smaller parties the Janata was in a position to muster a majority and form a stable government. However, the President refused to entertain Ram's proposal, ordered the dissolution of the House and asked Charan Singh to serve as caretaker Prime Minister until January when the  Limaye, The Collapse of the Janata Experiment, p. 515. Sharda Paul, 1980 General Elections in India (New Delhi: Associated Publishing House, 1980) p. 65. 105  106  59  elections could be held. In turn, the Janata Party announced that it would introduce an impeachment motion against the President in the Rajya Sabha. However, the President prorogued the upper house on August 25.  2.5  107  THE SEVENTH AND EIGHTH LOK SABHAS, 1980-89: CONGRESS DOMINANCE RESTORED The seventh general election resulted in a landslide victory by the Congress(I)  Party. As in previous elections, the CongressQ) did not form electoral alliances with parties at the national level, however, it engaged in a successful coordination of electoral strategy with selected political parties at the local level in different states. The most important of these Congress(I)-allied parties were the DMK in Tamil Nadu, the National Conference in Kashmir, the Moslem League and the Kerala Congress (Joseph Group) in Kerala.  108  109  The Congress(I)'s alliance with the DMK was particularly interesting and indicative of the party's renewed ability to build winning electoral coalitions. As mentioned earlier, the Congress had formed an electoral alliance with the DMK before during the 1971 general elections in Tamil Nadu. However, the relationship between the two parties soured after Prime Minister Indira Gandhi had the DMK state government removed and President's Rule imposed on Tamil Nadu in 1975. Thus, in the 1977 general  For discussions about the controversy that the President's decision invoked, see the following works by James Manor, "Seeking Greater Power and Constitutional Change: India's President and the Parliamentary Crisis of 1979" in D. A. Low, ed., Constitutional Heads and Political Crises: Commonwealth Episodes, 1945-85 (London, 1988) 41; "India" in D. Butler and D. A. Low, eds., Sovereigns and Surrogates: Constitutiona Heads of State in the Commonwealth (London, 1991), pp. 144-70; "The Prime Minister and the President" in James Manor, ed., Nehru to the Nineties: The Changing Office o Prime Minister in India (Vancouver, 1994), pp. 115-37. The Kerala Congress split into two factions in July 1979: the pro-Congress(I) Kerala Congress(Joseph) and the pro-Communist Kerala Congress (Mani). See, Keesing's Contemporary Archives (1980), p. 30255; and Political Parties in Asia and the Pa 345. 109 Keesing's Contemporary Archives (1980), pp. 30325-30327. 107  108  60  election the D M K allied with the Janata Party, however, their alliance failed badly as the Congress(Ruling)-ADMK  combine won 32 of the 39 seats that were contested in Tamil  Nadu. In the 1980 election, Mrs. Gandhi's Congress(I) and the D M K formed an alliance once again resulting in a spectacular performance in the state: of the 39 seats the Congress(I)-DMK alliance won 36! The Congress(I)'s alliance did not work out as well in Kerala where the triple alliance o f the Congress(I), the Moslem League and the Kerala Congress (Joseph Group), also called the United Democratic Front, won 8 of the states' 20 L o k Sabha seats while the remaining 12 seats were won by the Left Democratic Front including the Congress (Urs) , the Communist Party of India (Marxist), the Communist Party o f India and the 110  Kerala Congress (Mani Group). In the previous election of 1977, Mrs. Gandhi's CongressfRuling) performed significantly better in the state as it managed to win 11 o f Kerala's 20 seats. In Kashmir, the front of the Congress(I) and the National Conference won 4 o f the state's 6 L o k Sabha seats, although only 1 went to the Congress(I).  111  Overall, the Congress(I) won 42.7% o f the popular vote and 66.7% o f the L o k Sabha seats in the 1980 election. Including its electoral partners, the Congress(I)-alliance controlled 7 1 % o f the seats in the lower house of parliament. The Congress(Urs) performed miserably by winning a mere 13 seats.  112  O f the Janata fragments, the Charan  Singh-led Janata Party(Secular) outperformed the Janata Party by winning 41 seats as opposed to the latter's 31. Even though the Janata Party fielded a candidate in almost every constituency, unlike the Janata Party (Secular) which ran candidates in only 294 o f  110  M r . Devraj Urs left the Congress(I) shortly before the 1980 elections and became  president o f the Congress(Singh), which was the continuation o f the Congress(Reddy). The name change reflected the assumption o f the party's leadership by Swaran Singh. Once Urs joined and became president of the party, the Congress(Singh) was renamed  Congress(Urs). See, Political Parties  ofAsia and the Pacific, p. 337. onKeesing's Contemporary Archives (1980) pp. 30325-7, and Butler, R o y Lahiri, India Decides, various state tables. ln  Based  U2  I t is worth noting that at the time of the dissolution o f the L o k Sabha, the Congress(I)  had only 80 and the Congress(Urs) 56 seats.  and  61  the 529 electoral districts, and formed electoral alliances with important local parties, such as the ADMK in Tamil Nadu and the Akali Dal in Punjab, it fared miserably at the polls. The election was followed by the formation of a Congressfl) majority government with Indira Gandhi at its head on January 14, 1980. The massive defeat of both Janata fragments at the polls led to further splintering in their ranks. First, the Janata Party (Secular) broke up between followers of Charan Singh, the former Prime Minister, and Raj Narain, the founder of the party. The latter joined hands with H. N. Bahaguna, the leader of the CFD faction that defected from the Janata Party when the Janata Party (Secular) was formed in 1979, to form the Socialist Democratic Party. Second, the Janata Party suffered two separate splits as first its leader, Jagjivan Ram, and his faction left the party to join the CongressfUrs) , and then as the 113  Jana Sangh members quit it and formed the Bharatiya Janata Party (B JP) in December 1980.  114  The Congress(I) government remained firmly in power throughout the term of the Lok Sabha even surviving the terrible assassination of Prime Minister Indira Gandhi on October 31, 1984. Within just a few hours of her death the Congress(I) Parliamentary Board recommended to President Zail Singh to appoint Rajiv Gandhi, son of the slain Prime Minister, as the new head of government. The President complied with the advice  The leadership of the Janata Party was assumed by Chandra Shekhar after Jagjivan Ram's departure. ""The BJP was very keen on distancing itself from its predecessor, the BJS, and its radical Hindu militancy. To this end, the BJP adopted the following five basic commitments as its founding principles: Gandhian socialism, nationalism and national integration, democracy, positive secularism and value-based politics. This platform clearly suggested that the BJP was trying to capture the center of the political space by incorporating the traditional positions of both the left (socialism, secularism) and the right (nationalism and value-based politics the latter being a veiled reference to Hindu values). See, Yogendra K. Malik and V. B. Singh, Hindu Nationalists in India: The Rise of the Bharatiya Janata Party (New Delhi: Vistaar Publications, 1994), pp. 37-38; and Bharatiya Janata Party, Our Five Commitments (Delhi: Asiatic Printers). ,13  62  and appointed Rajiv Gandhi the same day. On November 2, the Congress(I) Parliamentary Party elected Rajiv Gandhi as its leader, and on November 12, the Congress(I) Working Committee along with Congress(I) Chief Ministers and state party presidents elected him as co-president of the party.  115  The eighth general election held in 1984 resulted in the best electoral performance of the Congress(I) ever both in terms of the share of the popular vote it mustered, 48.1%, as well as the share of Lok Sabha seats it won, 76%. As before, the Congress(L) refrained from entering electoral alliances on a comprehensive national scale, however, it did continue its tradition of local alliances with selected political parties. In yet another curious twist of relationships, the Congress(I) allied with the AIADMK in Tamil Nadu. The partners swept the state by winning 37 of 39 seats. Similarly, this time the Congress(I)-led United Democratic Front in Kerala also improved on its earlier performance: in contrast with the 8 seats that thefronthad won in 1980, it succeeded in winning 17 of the state's 20 seats in 1984! In fact, the Congress(I) Party gained more seats in every state of the country than it had in 1980 with the exception of Andhra Pradesh, the only state where it lost miserably, Gujarat, Karnataka, Punjab, and Dadra and Nagar Haveli.  116  The opposition was utterly devastated in the election. The largest opposition party emerging was the Telugu Desam (TD), a state party recognized and running candidates only in the state of Andhra Pradesh, with a mere 30 seats. Of the national parties , the 117  Communist Party of India (Marxist) won the most seats, 22 of the 64 that it contested. The proposed cabinet list that Rajiv Gandhi submitted to President Zail Singh reflected the young Prime Minister's attempt to establish firm control over the  The other president of the Congress(I) party was Mr. Kamlapathi Tripathi. Compared to its performance in 1980, the Congress(I) lost 35 seats in Andhra Pradesh, 1 in Gujarat, 3 in Karnataka, 6 in Punjab, and 1 in Dadra and Nagar Haveli. See, Butler, Roy and Lahiri, India Decides, pp. 110-111. For a definition see page 29, footnote 44. 115  116  U7  63  government as more than half of his proposed cabinet consisted of newcomers. While a majority of his ministers with cabinet rank were incumbents, at the rank of ministers of state he nominated 16 newcomers while retaining only 9 incumbents. In a further 118  attempt at establishing his authority, Rajiv Gandhi secured the passage of the Fifty Second Amendment to the Constitution, which laid down very strict rules about defections, splits and mergers in the legislature. The novice Prime Minister was keen 119  Of the 15 ministers with cabinet rank, there were six novices: Abdul Gafoor in the Works and Housing portfolio, Ashoke Sen in Law and Justice, Bansi Lai in Railways, H. K. L. Bhagat in Parliamentary Affairs, and K. C. Pant in Education. See, Keesing's Contemporary Archives (1985), p. 33466. This Amendment established the Tenth Schedule of the Constitution which specified in paragraph 2 that a member of a legislature, state or national, belonging to any political party shall be disqualified for being a member (lb) if he votes or abstains from voting in such House contrary to any direction issued by the political to which he belongs or by any person or authority authorized by it in his behalf, without obtaining, in either case, the prior permission of such political party, person or authority and such voting or abstention has not been condoned by such political party, person or authority within fifteen days from the date of such voting or abstention; 118  119  (2) An elected member of a House who has been elected as such otherwise than as a candidate set up by any political party shall be disqualified for being a member of the House if he joins any political party after such election. Furthermore, paragraph 3 of the Schedule states that Where a member of a House makes a claim that he and any other members of his legislature party constitute the group representing a faction which has arisen as a result of a split in his original political party and such a group consists of not less than one-third of the members of such legislature party, (a) he shall not be disqualified under sub-paragraph (1) of paragraph 2 ... (b)fromthe time of such split, such faction shall be deemed to be the political party to which he belongs for the purposes of sub-paragraph (1) of paragraph 2 and to be his original political party for the purposes of this paragraph. Finally, according to paragraph 4 of the Schedule, (1) a member of a House shall not be disqualified under subparagraph (1) of paragraph 2 where his original political party merges with another political party and he claims that he and any other members of his original political party (a) have become members of such other political party or, as the case may be, of a new political party formed by such merger; or (b) have not accepted the merger and opted to function as a separate group ...  64  on ensuring that party discipline would be strictly observed and enforced during his tenure and that he would not have to face the situation that his mother, Indira Gandhi, did when the governing party split in 1969.  120  Interestingly, even though the non-Congress parties were never so weak in the Lok Sabha as they were after the 1984 election, it was during the term of this parliament that they managed to establish a cooperative framework that would leave a much longer lasting impact on party politics than the brief interlude of the Janata Party did. The revival of the parliamentary opposition originated in conflicts within the Congressfl) Party, similarly to the process that had started with the split of the Congress party in 1969 and culminated in the eventual creation of the Janata Party in 1977. This time, the protagonist was V. P. Singh, Rajiv Gandhi's Finance Minister. Singh was 121  given a broad mandate by the Prime Minister to eliminate corruption and cleanse the economy. However, when Singh got engaged in probing the illicit finances of individuals with close connections to the Prime Minister, he was quickly re-assigned to the Defense portfolio. As Defense Minister, Singh appeared to continue his crusade against corruption when he ordered an inquiry into alleged irregularities committed in the government's purchase of two submarines from Germany on April 9, 1987. The Prime Minister  (2) For the purposes of sub-paragraph (1) of this paragraph, the merger of the original political party of a member of a House shall be deemed to have taken place if, and only if, not less than two-thirds of the members of the legislature party concerned have agreed to such merger. Hardgrave and Kochanek note that there have been "more than 2,700 recorded cases since 1967, most within the state assemblies. Congress had been the principal beneficiary, with as many as 1,900 defections to Congress. ... Between March 1967 and June 1968, the high days of defection, 16 state governments were brought down by defections. Of the 438 legislators who changed parties during this period, 210 were rewarded with ministerships." See, Hardgrave and Kochanek, India: Government and Politics, p.273. As shown earlier, defections played an important role in the destabilization of the Congress government in the Lok Sabha in 1969 as well as in the collapse of both the Janata Party-Akali Dal and the Janata Party(Secular)Congress(Reddy) governments.. 120  121  See, Hardgrave and Kochanek, India: Government and Politics, pp. 276-7.  65  denounced Singh's effort and, as a result, the Defense Minister resignedfromthe cabinet on April 12. Shortly thereafter he was officially expelledfromthe Congress(I) Party along with three other former ministers who had also resignedfromtheir posts.  122  Following Singh's departurefromthe Congress(I), eleven opposition parties met in New Delhi to deliberate about a coordinate position against the government. While the meeting succeeded in drafting a joint appeal to fight corruption in government and to demand a parliamentary inquiry in the government's defense dealings, it failed to accept a proposal that was put forth to create a confederation of anti-Congress(I) parties. Later 123  on in September, ten opposition parties met at Surajkund, Haryana, where they resolved to start a joint campaign demanding the resignation of the Congress(I) government and the call of early elections. The ten parties attending the meeting were the Lok Dal (Bahuguna), the Telugu Desam, the BJP, the Janata Party, the Congress(S), the Asom Gona Parishad, the Akali Dal (Badal), the DMK, the National Samajwadi Congress, and the Nagaland National Party. Noticeably, none of the left parties were invited to the 124  meeting indicating that similarly to the process of opposition consolidation in 1975-77, the Communist parties were no less liked by most opposition parties than the Congress(F) itself. In particular, the relationship between the Communist parties and the BJP was especially strained. Whereas the BJP had always considered the Communist Parties antinational, the Communist parties rejected the BJP's Hindu nationalism, no matter how veiled it was. A recent report of the BJP Working Group surveying party members' opinion about the party's relationship with other parties had concluded the following: the party must never merge with any other party; the BJP must fight the election on its own These were Arun Nehru, former Rajiv confidante and Minister of State for Defense, V. C. Shukla, and Arif Mohammed Khan, Minister of State for Industry and Company Affairs. The wave of ministerial departures continued further as Mufti Mohammed Sayeed, Minister of State for Tourism, also resigned in July 1987. Asian Recorder (1987), 19544. Ibid, 19728. 122  m  UA  66  manifesto and its own symbol; and finally the BJP  must never have any truck with the  two Communist parties and the Muslim League. However, party members were in favor of coordinating candidate nomination with other opposition parties on the basis of mutuality and reciprocity. Similarly, shortly before the 1989 election, the BJP reiterated 125  its position regarding the Communist parties in a resolution of the party's National Executive: "The ouster of this corrupt and worthless Government is the No. 1 issue before the people and should take precedence before everything else. The BJP is prepared to co-operate with all democratic and nationalist [read non-Communist] parties on this issue [italics added]."  126  The two Communist parties were no less antagonistic against the BJP. The electoral manifestos that the two parties issued before the 1989 election harshly condemned both the BJP and the ruling Congress(I) for stirring communal conflicts all across the country. The CPI(M) manifesto in 1989 clearly identified that the party had two objectives in the election: to defeat the Congress(I) Party and isolate the communal forces [read BJP]. The CPI also declared that its objective was "the end of the rule of Mr. Rajiv Gandhi and the Congress(I), the defeat of communal and secessionist forces and the preservation of secular democracy." Furthermore, the CPI also stated that it would oppose any government of which the BJP was a partner.  127  It was the traditional policy of the BJP, and its predecessor the BJS, not have any sort of cooperation with the Communists under any conditions. For instance, in 1966 the General Council of the BJS, the predecessor of the BJP banned "any sort of alliance with the Communist Parties, which have nothing in common with the Jana Sangh and which misuse the democratic process only for uprooting democracy." Bharatiya Jana Sangh, Party Documents (New Delhi: BJS central Office, 1973), volume 4, p. 109 Bharatiya Jana Sangh, Resolutions of the National Executive Meeting in Ahmedab 125  126  October 7-9, 1988, pp. 15-6. The Statesman (November 1, 1989)  l21  p. 7. Jyoti Basu, leader of the Communist Party of India (Marxist) and Chief Minister of West Bengal, went on record at a rally held in New Delhi in December 1987 claiming that to remove the Rajiv Gandhi government "we shall have to unite democratic and Leftist forces. All parties, except the communal BJP can come together. Even those coming out of the Congress(I) are welcome. We are ready to provide leadership." Asian Recorder (1987), 19833.  67  A few days after the conclusion of the Surajkund conclave, prominent exCongress(I) leaders who had been recently expelled from the party launched the Jan Morcha, an allegedly non-political forum, under V. P. Singh's leadership on October 2, 1987.  128  The efforts of the opposition were revealed by the Rajya Sabha election in March  1988. Reflecting the growing strength of the opposition at the state level , the 129  Congress(I) lost seven seats in the upper house bringing the total number of its seats to 141. In addition, the largest of the Congress(I)'s allies in the upper house, the A I A D M K , split and the larger of the two successor parties sided with the opposition. Thus, the 130  Congress(I) had to create more alliances in the upper house so that it would have the twothirds majority required for the passage of any constitutional amendment bill that the government would seek to have passed. In August 1988 the Telugu Desam, the Janata Party, the Congress(S), the D M K , the AGP and the Janata Morcha agreed to form a federation called Rashtriya Morcha, or National Front under the leadership of V. P. Singh and N . T. Rama Rao, the leader of the Telugu Desam.  131  The participating parties agreed that the AGP, the D M K and TDP  would be recognized as the dominant parties of the Front in the states of Assam, Tamil Nadu and Andhra Pradesh respectively. It was in these states that these parties had their electoral support concentrated and formed state government alternating with the Congress(I) and its allies. Shortly thereafter, the Jan Morcha merged with the Janata  *Asian Recorder (1987), 19728.  l  0 f the major states, the Congress(I) was in opposition in Andhra Pradesh, Assam, Haryana, Karnataka, Kerala, Punjab, Tamil Nadu, and West Bengal. This was particularly important because seats in the Rajya Sabha arefilledby the representatives of the states who are chosen by the respective State Assemblies using the single transferable vote formula, as specified in paragraph 80 and the Fourth Schedule of the Constitution. 1 2 9  The A I A D M K split between those who followed the wife of the party's deceased leader and those who sided with his mistress. Eventually, the latter would become the more prominent of the two. ™Asian Recorder (1988), p.20171-2. 130  68  Party as well as the warring factions of the Lok Dal to become the Janata Dal.  132  The creation of the Janata Dal and the National Front signaled the entry of a potentially very strong new player in the party system that could challenge the Congress(D's in two ways. First, being an amalgam of the non-Congress(I) centrist national parties, the Janata Dal challenged to invade the Congress(I)'s position in the ideological space of the party system. Second, by forming an alliance with some of the most formidable local opponents of the Congress(I), the Janata Dal could also destabilize the Congress(I)'s electoral bases in those areas where its centrist appeal might otherwise not yield the desired benefits. The National Front parties agreed to run on a common platform in the next general election and continue their cooperation if elected to power. Thus, a clear commitment was made to the possibility of forming a coalition government in the aftermath of the polls. However, in order to ensure the defeat of the Congress(I), the National Front had to arrive at an electoral understanding with the Communist-led Left Front, and the BJP on the right. In the end, this effort failed as neither the Left nor the Right wanted to cooperate with one another in the election. This meant, that the National Front had to engage in separate coordination with the two sides respectively, which, however, only weakened it because it had to make compromises in two different directions. Nonetheless, the opposition parties had made huge progress in terms of consolidating their ranks and offering a potential alternative to the Congress(I). In addition, it must also be noted that the opposition parties were very successful in coordinating their strategies in the two houses of parliament as the term of the eighth Lok Sabha was drawing to its close. In July 1988, fifteen parties, including both Communist parties as well as the BJP!, signed a memorandum of understanding to oppose the Congress(I) in a concerted fashion. After organizing a number of demonstrative walk-outs from both houses, the opposition  1  Asian Recorder (1988), p. 20299-302.  69  parties managed to defeat the government-sponsored Panchayati Raj and Nagarpalika Bills, both of them requiring an amendment of the constitution and therefore a two-thirds majority in both houses.  2.6  133  T H E NINTH L O K SABHA, 1989-91: T H E J A N A T A F A I L U R E R E P E A T E D The 1989 general election was the first ever in India's electoral history that  produced a hung parliament. Although the Congress(I) Party remained the single largest party in the Lok Sabha it only held 36.3% of the seats. As in earlier elections, the Congress(I) formed alliances with parties at the local level, most importantly with the AIADMK(Jayalalitha) in Tamil Nadu and the Moslem League and the Kerala Congress (Mani) in Kerala. In both states, the Congress(I)-led alliance did very well: the Congress(I)-AIADMK grabbed all but one of the seats in the state, while the United Democratic Front won all but three seats in Kerala.  134  The National Front arrived at a perfect coordination of its members by determining which one of them would stand the best chance to defeat the Congress(I) in each district and ensuring that other members of the alliance would not run in that district. In total, the National Front ran candidates in 315 of the 528  13S  districts making  adjustments with the BJP and the Left Front in the remaining 213. Whereas the Left and the BJP had no electoral understanding whatsoever and they each ran candidates wherever they believed they would have a chance of either winning or at least hurting the other's chances of doing so, the Janata Dal and the BJP achieved a remarkable degree of cooperation: of the 252 seats in the Northern Hindi belt where both parties were very  "The Panchayati Revival," Seminar, 360 (August 1989). The Congress(I) won 27 seats, the A I A D M K won 11 seats in Tamil Nadu. In Kerala, the Congress(I) won 14, the Moslem League 2, and the Kerala Congress(Mani) 1 seat. See, Butler, Roy and Lahiri, India Decides, pp. 100-101. Although the total number of seats in the Lok Sabha was 543, elections were not held in Assam at the time. Therefore, only 529 seats were contested in 1989. 133  134  13S  70  strong they agreed to field a joint candidate in all but 21 of them!'  36  In total, the three opposition blocs, i.e. the National Front, the Left Front and the BJP, fielded an average of 1.2 candidates in each electoral district in the 1989 election. In stark contrast, the same parties, or their predecessors, had fielded an average 1.6 candidates per district in 1984 and 2.1 in 1980. In 1977, these predecessors of these parties had run an average 1.1 candidate per district. This shows that the opposition 137  parties were quite successful in recreating the electoral coalition that brought defeat to the Congress(I) in 1977. Of course, the big difference between the two elections is that the 1977 electoral coalition was much more concentrated than the one formed in 1989. Whereas in 1977 there was one party, the Janata contesting on the B L D symbol, opposing the Congress(I) in 75% of the districts, in 1989 the opposition party that competed against the Congress(I) in the most district was the Janata Dal fielding a total of 243 candidates. This fragmented nature of the electoral coalition that the opposition parties formed clearly suggested that unless the Congress(I) won a majority of the seats, the ninth Lok Sabha would be a hung parliament simply because none of the opposition parties ran the sufficient number of candidates that would have given it at least a mathematical possibility to win a majority. Indeed, the ninth Lok Sabha was a hung parliament. While there was no clear winner in the election in the absolute sense of the term, the Janata Dal was the most successful national party in terms of the rate of its candidates' victory: 58% of the Dai's candidates were elected in contrast with just 39% of the Congress(I), 38% of the BJP, 24% of the CPI and 51% of the CPI(M) candidates. Interestingly, the regional, or state-, parties that belonged to the opposition camp scored a major failure. The D M K ' s victory  ""Statesman (November 4, 1989), p.l. The actual figures are as follows: 656 candidates for the 529 in 1989; 875 candidates in 541 districts in 1984; 1092 candidates in 529 districts in 1980; and 594 candidates in 543 districts in 1977. 137  71  rate was 0 %  1 3 8  and the Telugu Desam's 6%. In contrast, the Congress(I)'s regional allies  performed very well: the AIADMK, the Moslem League and the Kerala Congress(Mani) had a victory rate of 100% each.  139  Concurrently with the national election in November 1989, the states of Andhra Pradesh, Karnataka and Uttar Pradesh also held State Assembly elections. The opposition parties succeeded in defeating the Congress(I) and forming a non-Congress government only in Uttar Pradesh where the election produced a hung assembly with the Janata Dal being the single largest party. In the other two states the Congress(I) managed to win comfortable majorities in the Vidhan Sabha. Not surprisingly, the Congress(I) also won the overwhelming majority of the Lok Sabha seats in Karnataka and Andhra Pradesh. Following parliamentary convention, the President of the Republic, R. Venkataraman, invited Rajiv Gandhi, the leader of the largest party, to form a government. However, the Congress(I) declined the invitation, arguing that the electorate had clearly spoken and wanted a change in government. However, there were also 140  voices suggesting that the Congress(I) leadership expected differences among the opposition to surface any time soon rendering any non-Congress government that might form to be potentially unstable. The Congress(I) would then be in a more advantageous position and could even form a government on its own once it faced a divided opposition again. Following the Congress(I)'s decision to refrain from forming a government, the National Front appointed a committee to solicit and negotiate the support of the BJP and the Left for a National Front government. The committee comprised N . T. Rama Rao, leader of the Telugu Desam Party, V . P. Singh, the leader of the Janata Dal, Devi Lai,  Not one of the party's 31 candidates was successful in 1989. Butler, Roy and Lahiri,India Decides, pp. 100-101. Sharda Paul, 1989 General Election in India (New Associated Publishing House 1990), p. 54.  138  139  140  72  Ajit Singh and Arun Nehru, each being a senior Janata Dal official.  141  While the Left  appeared to be quite ready to provide support to the government from the outside, i.e. without formally taking up any ministerial portfolios, the BJP expressed two concerns. As L K . Advani, the president of the BJP wrote to Rao and Singh,  (i) The National Front and the BJP fought these elections on two separate manifestos, not on a common manifesto. A manifesto is a party's solemn commitment to the people. Our two manifestoes have several common features, such as grant of autonomy to Akashvani and Doordarshan, enactment of a Right to Information Act, incorporating Right to Work as a fundamental right in the Constitution, elimination of corruption by the creation of an institutional watchdog like the Lokpal, taking steps to give debt relief and ensure remunerative prices to the farmer, etc. But there are aspects on which the two manifestoes differ. We would like the N. F. government to confine its governmental programme to issues on which we agree. (ii) The main constituent of the National Front is the Janata Dal. Ever since its launching, J. D. leadership, by its utterances and actions, has been consciously trying to convey to the people an impression that it regards the BJP as a communal party, and that it would rather sit in the opposition than ever share power with it. The J. D.'s public postures have thwarted the building up of any abiding relationship of trust and friendship between our parties. If it is acknowledged by the J. D. that though the J. D . and BJP differ on issues like Art. 307, Uniform Civil Code, Human Rights Commission, Ram Janmabhoomi, etc., the J. D. does not regard the BJP as communal, that would go a long way in removing misgivings in our rank and file. I hope the N . F. will take note of these reservations and exert to obviate them. The BJP is keen to see that the Ninth General Election marks the end of Congress rule in New Delhi. It is, therefore, that even while expressing these reservations, we have not made our support to you conditional to your agreeing to remove them. In response to your letter, the BJP wishes to convey to you its readiness to give general but critical support to theN. F. government. 142  Having thus secured the support of both the BJP and the Left, the National Front proceeded to form a government. Upon swearing in the new cabinet the President indicated that the new Prime Minister, V . P. Singh would have to prove within one month that his government enjoyed majority support in the Lok Sabha.  143  The new government  included the Janata Dal and its tiny National Front partners each of which received one  Paul, 1989 General Election 'Ibid., pp. 55-6. 'Ibid., p. 59.  in India,  p. 55.  73  cabinet portfolio  144  T A B L E 2.7: E L E C T O R A L COORDINATION B E T W E E N T H E BJP A N D T H E J A N A T A D A L IN 1989 A N D 1990  % of votes Seats in 1989 Lok Seats won in Seats Sabha contested in 1990 contested in Seats won election, 1990 Vidhan 1989 Lok in 1989 Lok BJPrJD Vidhan Sabha Sabha poll, Sabha poll, Sabha poll, poll,  State  BJPrJD  BJPrJD  Arunachal Pradesh Bihar  22(2)  ooo  2.0%:1.6%  2537 (54)  931  13%36.4%  Gujarat  12:14 (26)  12:11  Hirnachal Pradesh  42(4)  300  Madhya Pradesh  33:11(40)  27:4  Maharashtra 3323 (48)  1005  21:19(21) 17:13 (25)  006 13:11  Orissa Rajasthan  Pondicherry N A ( 1 )  NA  % of votes in 1990 Vidhan Sabha poll, BJPrJD  BJPrJD  BJPrJD  0:52 (60)  0:10  033.3%  39:123  11.6%25.7  67:70  26.7%29.4  46:11  41.8%:10.8  237277 (324) 30.5°/o27.8% 143:147 (182) 45.3%:7.2% 51:17(68)  39.7%:8.3% 269:115 22028 (320) 23.7%:11% 104214 4224 (288) 9.5%34.6% 63:139(147) 2:123 29.6%:25.7% 128:120 85:55 (200) NA 8:8 (30) 0:4  39.1%:7.7% 10.7%:12.7 3.6%53.7% 25.3%:1.6% 0.5%:0.1%  Source: Butler et al, various state tables. Note: The numbers in the brackets in the second and fifth columns indicate the number of Lok Sabha and Vidhan Sabha seats respectively available in the given state. Soon after the formation of the government, subnational elections were due to be held in Arunachal Pradesh, Bihar, Gujarat, Hirnachal Pradesh, Madhya Pradesh, Maharashtra, Manipur, Orissa, Pondicherry and Rajasthan. The former opposition parties  144  Although the A G P had no representatives in the Lok Sabha, given that there were no elections held in Assam, one of the party's representatives in the Rayja Sabha was included in the cabinet.  74  did very well in each of these states in the 1989 parliamentary election with the exception of two tiny states of Arunachal Pradesh and Manipur, where the Congress(I) grabbed both seats that it contested in each, Pondicherry, where the Congress(I) won the single seat contested, and Maharashtra, where the Congress(I) won a majority of the 48 seats. As Table 2.7 illustrates, the BJP outperformed both the National Front and the Left in Hirnachal Pradesh, Madhya Pradesh, and Maharashtra both in terms of the share of the popular vote it polled as well as the number of seats it won. On the other hand, the Janata Dal performed the best of the three allied blocs in Bihar and Orissa. In the states of Gujarat and Rajasthan the BJP and the Janata Dal won almost the same number of seats although in both states the BJP's share of the popular vote was slightly higher than that of the Janata Dal. The Left commanded only a negligible presence in any of these states. In exchange for its support of the National Front government in the Lok Sabha, the BJP expected the Front to help it win the elections, by agreeing to electoral alliances, and form state governments in the states where it had done so well in the parliamentary election.  145  In four states, however, this cooperation was very difficult to achieve. In  Maharashtra, the BJP was already committed to cooperating with a local party, the Shiv Sena, known for its virulent position against the Left. By agreeing to support the BJPShiv Sena alliance, the National Front would have risked to lose the support of the Left Front in the Lok Sabha. In Orissa, the local Janata Dal unit was unwilling to cooperate with the BJP, however, the BJP did not see it worthwhile fielding the maximum number of candidates in the state given its negligible electoral base there. In Gujarat and Rajasthan, the problem that the parties ran into was that they both performed very well in the Lok Sabha polls in these states. Thus, when it came to negotiating the electoral alliance for the 1990 Vidhan Sabha election, both parties claimed that they could have performed even better had they not had a seat adjustment with the other. Sunday (January 21-27, 1990), p. 29. ^Sunday (January 28-February 3, 1990), pp. 30-1. U5  146  The BJP  75  further stressed that it was now the Janata Dai's turn to reciprocate the BJP's assistance to the national government in the Lok Sabha. This argument, of course, did not appeal to the local units of the Janata Dal in these states. Eventually, the talks on alliance building in Gujarat and Rajasthan broke down and the two parties ran candidates on their own. Following the elections, the Janata Dal formed a minority government in Bihar with the external legislative support of the BJP, while the BJP formed minority governments in Gujarat, Rajasthan, Madhya Pradesh, and Himachal Pradesh. The relationship between the Janata Dal and the BJP was far from amicable at either the national or the subnational levels. In particular, the major issue that divided the two parties was the disputed construction of a Hindu temple at Ayodhya, the birth place of Lord Ram, a Hindu deity, on the site of the Babri Masjid mosque, a revered site of the Muslims. The BJP had supported the temple construction movement led by the Vishwa Hindu Parishad (the World Hindu Federation) that had set October 30, 1990 as the date when the demolition of the mosque and the construction of the temple should commence. The President of the BJP, Lai Krishna Advani announced in July that  The BJP will participate in full strength in any agitational program that might be launched by the VHP for the construction of the temple. Any attempt at settling the VHP's plans will snowball into 'the greatest mass movement this country has ever witnessed. We will see to it that with the support of nationalist forces, the Sri Ram temple is constructed. 147  The BJP's position was especially troubling for the Janata Dal government in Uttar Pradesh, the state where the disputed site was situated, because it received substantial electoral support from the Muslim community in the previous election, as shown in Table 2.8, which had to be reciprocated by his government's protecting the community's interests against the agitation of Hindu fundamentalists.  U 1  Sunday (July 15, 1990), p. 42.  76 T A B L E 2.8: N U M B E R A N D P E R C E N T A G E OF SEATS W O N B Y P A R T Y B Y M U S L I M P O P U L A T I O N PROPORTION I N T H E 1989 L O K S A B H A E L E C T I O N S I N U T T A R  Janata Dal Congress(I) BJP  Muslim population 20-50% 16 (66.7%) 3 (12.5%) 3 (12.5%)  Muslim population > 50% 34 (59.6%) 12 (21.1%) 5 (8.8%)  50 (61.7%) i ' 15 (18.5%) —i ' 8 (9.9%)  2 (8.4%) 24 (100.1%)  6 (10.5%) 57 (100%)  8 (9.9%) i L 81 (100%)  Others Total  Note: Bracketed entries re"er to Dercentace nf urate ,  — v «  e  w  ~* ovuu.  T  i i . v u a u v v c i t u u 5 M i i g IUI IOUT  Total  oi me 8 3 parliamentary  districts. Source: Paul R. Brass, "Caste, Class, and Community in the Ninth General Elections for the Lok Sabha in Uttar Pradesh," in Harold A. Gould and Sumit Ganguly eds., India Votes: Alliance Politics and Minority Governments in the Ninth and Tenth General Elections O^oulder: Westview Press, 1993), p. 134. Under pressure from Mulayam Singh Yadav, the Chief Minister of Uttar Pradesh, Prime Minister V . P. Singh announced on July 28, 1990 that the V H P ' s unilateral deadline of October 30 would not be tolerated by the national government. The BJP finally decided to withdraw its support from the Janata Dal government in Uttar Pradesh, which, however managed to remain in office by mustering additional support from small parties and Independents.  148  The second contentious issue which soured the relationship between the Janata Dal and the BJP was concerning the issue of reserving national government jobs for members of the so-called Other Backward Castes. When pledging its support to the National Front after the election, the BJP expected that the government would consult it on major policy issues. However, in August 1990 the Prime Minister announced, without prior consultation with either the Left or the BJP, that his government decided to implement the recommendations of the Mandal Commission by seeking to reserve 27% of all central government jobs to members of the Other Backward Castes on top of the  '"Sunday (June 8, 1990), p. 12.  77  22.5% that was already reserved constitutionally for members of Scheduled Castes and Scheduled Tribes.  149  The sudden announcement of the new reservation policy upset the BJP not only because it was not consulted but, more importantly, because, of the potential impact of this policy on its own electoral prospects in the future. Since the forging of a pan-Hindu unity transcending the multifarious cleavages dividing the community was central to the center of the BJP's political and electoral strategy, the National Front government's reservation policy explicitly seeking to divide the Hindu community along caste lines by pursuing the social, economic and political empowerment of the backward castes, posed a direct challenge to it. Apart from the temple and the reservation issues, the BJP was also apprehensive about what it perceived to be the appeasement of the Muslim community by the National Front government. The declaration of the prophet Mohammed's birthday as a public holiday; the Prime Minister'sfrequentconsultations with the Shahi Imam of the Jama Masjid of Delhi; the dismissal of Jagmohan, the governor of Jammu and Kashmir who was cracking down on Muslim fundamentalists in that state; and the failure to stand up decisively against Pakistan's interference in the affairs of Kashmir and Punjab were all issues that the BJP complained about.  150  In immediate response to the announcement of the Mandal policy, the BJP  It is worth noting that this announcement came shortly after the Prime Minister had removed his Deputy Prime Minister, Devi Lai, from the cabinet in July over allegations of the latter's involvement in electoral misconduct in the state assembly elections of Haryana state. Devi Lai had been at odds with Prime Minister V.P. Singh ever since the latter had defeated him in the race for the National Front's leadership right after the elections. Since Lai was widely regarded as a prominent leader among the Other Backward Castes the new reservation policy was intended to cut into his politicalelectoral base and weaken his political position against the Prime Minister. India Today (September 15, 1990), pp. 31-2. 149  'Sunday (July 25, 1990), p.68.  78  executive resolved to launch a rathyatra \  a march through the country, to reach  ]5  Ayodhya by October 30 and begin the construction of the temple there. The BJP leadership concluded that the temple issue would unite Hindus regardless of caste and class divisions that the Mandal policy was aiming to institutionalize.  152  As October 30 was approaching the Prime Minister maintained that the High Court's order to uphold the status quo would be enforced. When Lai Krishna Advani and his entourage were about to enter the state of Uttar Pradesh in the final stage of their yatra, they were arrested by the police force of the state of Bihar on October 23. The BJP immediately notified the President of the Republic that the party no longer supported the National Front government in office.  153  In turn, the President instructed V . P. Singh to  seek a vote of confidence in the Lok Sabha on November 7. Two days before the vote was to be taken, on November 5, the Janata Dal split between the factions loyal to the Prime Minister and those following former Deputy Premier Devi Lai and his ally, Chandra Shekhar. The dissidents, who had the support of 154  about 60 Janata Dal members of Parliament, formed a new party, called the Samajwadi Janata Dal, or Janata Dal Socialist, and issued a statement demanding the Prime Minister's resignation. In the meanwhile, the Congress(I) expressed its support for the new party and stated that it was willing to support "the Janata Dal minus V . P. Singh" in office.  155  On November 7, with the BJP, the Congress(I) and the Samajwadi, as well as  some other parties, voting against it, the National Front government lost the confidence vote by 346 against and only 142 in favor.  156  Following the collapse of the National Front cabinet, the Congress(I) was again The phrase literally means "pilgrimage by chariot". Y . Malik and V. B. Singh, Hindu Nationalists 1990), p. 35.  151  152  in India,  p. 87; India  Ibid, p. 88.  ia  See footnote 114 for a brief description of the origin of this factional conflict. Frontline (November 10-23, 1990), p. 14. li6 Frontline (November 10-23, 1990), p. 6. 154  155  Today (October 15,  79  invited by the President to form a government by virtue of its still being the largest party in parliament. However, once more the party declined to do so. Instead, the Congress(I) pledged its external support to a government to be formed by the Samajwadi Janata. The Congress(I) made its offer credible by having extended support to the minority Janata Dal state governments in Gujarat, Haryana and Uttar Pradesh, whose hold on office became precarious after the BJP had severed its ties with the Janata Dal at all levels of government.  157  Following the Congress(I)'s declination of the President's invitation to form a government, the President duly approached the numerically next largest parties, the BJP and then the CPI(M) to see if they were interested or able to form a government.  158  However, both parties, and the blocs they led, unanimously refrained from accepting the invitation indicating that they would prefer the dissolution of the house and the holding of new elections. At that moment, Chandra Shekhar proposed that his tiny Samajwadi Janata Party would be able to form a government with the assistance of the Congress(I) and its allies. The President accepted Chandra Shekar's proposal, swore in both him as Prime Minister as well as his cabinet ministers on November 10 and instructed him to have his cabinet prove its majority within a couple of weeks in the House. With the assistance of the Congress(I), the minority Samajwadi Janata Party government won the vote of confidence on November 16, 1990 with 280 members of the Lok Sabha voting in favor, 214 voting against the new government while 11 abstained and 17 were absent.  159  The Congress(I) wasted no time trying to exploit its position by forcing Samajwadi leaders in Gujarat, Haryana and Uttar Pradesh, whose governments it was supporting, to pressure their party's national leadership to fall in line with the  The BJP had withdrawn its support to the Janata Dal governments in Haryana and Uttar Pradesh some time before it had done so at the national level. See, Sunday (July 29, 1990), p. 14; also Frontline (November 24-December 7, 1990), pp. 14-5. X5 *Frontline (November 24-December 7, 1990), p. 9. ^Frontline (November 24-December 7, 1990), p.5. 157  80  Congress(I)'s demands and preferences in the national legislature. This, of course, led to tension between the Samajwadi Janata Party's national and subnational leadership as the former sought to maintain its autonomy from the Congress(I) Party's dictates as much as possible. For example, the Chandra Shekar government authorized the use of Indian airports by American warplanes during the Gulf War, ignoring the protest of the Congress(I) Party that did not want to hurt the sentiments of the Muslim electorate in the country. However, the cabinet could not exercise such autonomy on most decisions. 160  Thus, for example, it had to obey the Congress(I)'s instructions to dismiss the government and impose President's Rule on the state of Tamil Nadu even though the Governor of the state had not filed a request to this end with the central government, as the Constitution required.  161  With time the state units of the Samajwadi Janata became increasingly weary of the Congress(I) and its alleged attempts at de-stabilizing their governments. Fearing that the Congress(I) would instigate dissidence within its ranks, the Samajwadi government of Haryana ordered plainclothes policemen to maintain round-a-clock surveillance on the residence of Rajiv Gandhi, President of the Congress(I) Party, in order to identify Samajwadi legislators from Haryana who might be contacting the Congress(I) leader. Having found out about this, Rajiv Gandhi demanded the dismissal of Om Prakash Chauthala, the Chief Minister of Haryana and general secretary of the national Samajwadi Janata Party.  162  To add weight to the demand, Gandhi ordered the suspension  of his party's support to the Samajwadi government in the national legislature until the demand was met.  163  Since all this happened as the vote was about to be taken on the  ''"Frontline (March 16-29, 1991), p . l l . At the time, the D M K , a member of the National Front, was in office in Tamil Nadu. The Congress(I) used its leverage at the national level to please its local ally, the A I A D M K in Tamil Nadu which was the main opposition party to the D M K government. In addition, Chauthala was also the son of Devi Lai, the leader of the rebel faction of the Janata Dal that came to form the Samajwadi. '"Sunday (March 17-23, 1991), p. 32. 161  162  81  motion of thanks to the President's speech, a vote classified as a matter of confidence, the very survival of the Chandra Shekhar government was at stake.  164  The Prime Minister  decided not to bow to the Congress(I) Party, advised the President to dissolve the house and to order new elections and subsequently resigned from his post on March 6, 1991. The President accepted Chandra Shekhar's resignation and asked him to remain in office in a caretaker capacity until the new elections were over.  2.7  T H E T E N T H L O K SABHA, 1991-96: A S T A B L E MINORITY GOVERNMENT In the 1991 election the National and Left Fronts coordinated their candidate  nominations just as they had done two years before. However, the BJP and the Congress(I) contested on their own, forming only local alliances with their respective regional supporters. Although the Congress(I) won 35 seats more than it had in the 1989 elections, it failed to win a majority of the seats in a number of large states including Bihar, Gujarat, Uttar Pradesh and West Bengal. Nevertheless, the party still emerged from the election as the single largest party in the Lok Sabha. The biggest gains in the election were made by the B P . In contrast with the 86 seats it had won in the previous election, in 1991 it managed to win 120. However, the BJP's performance did not mean that the party was gaining increasing support on a nation-wide scale. Quite the contrary, the BJP's gains were made almost exclusively in Uttar Pradesh where the party won 51 of the states's 85 seats in 1991 compared against the mere 8 seats it had won in the state in 1989. In all other states, except for Gujarat and Delhi, the BJP actually won fewer seats in 1991 than it had in 1989.  165  An important event during the election campaign was the assassination of former Prime Minister Rajiv Gandhi, the leader of the Congress(I), on May 21, the day after the Frontline (March 16-29, 1991), p. 11. Butler, Lahiri and Roy, India Decides, pp. 114-5.  XM  165  82  first of the three scheduled days of voting was concluded. Prior to his assassination, the Congress(L) was heading towards an electoral defeat that seemed to be worse than the one it had suffered in 1977. In the first phase of the election, held prior to Rajiv Gandhi's death, the Congress(I) Party won a mere 32% of the votes and only 25% of the seats.  166  After his death, however, the Congress did remarkably well by winning 39.3% of the votes and 62.2% of the seats that were contested in the post-assassination phases of the election. As Table 2.9 shows, the fortunes of the Congress(I) revived considerably vis-avis those of the other main parties after Rajiv Gandhi's assassination. The assassination, as Rudolph argues, was a major event because by "reframing the meaning of the campaign and the choices it offered" it influenced undecided marginal voters who in 167  turn could generate a counter-wave of sympathy in favor of the Congress Party. T A B L E 2.9: T H E IMPACT OF RAJIV GANDHI'S ASSASSINATION O N T H E E L E C T O R A L P E R F O R M A N C E OF T H E MAJOR PARTIES IN T H E 1991 L O K S A B H A ELECTIONS % of candidates % of candidates % of vote elected before elected after before % of vote after Party assassination assassination assassination assassination Congressfl) 24.80% 62.20% 31.90% 39.30% BJP 27.30% 40.10% 22.90% 18.50% Janata Dal 28.30% 12.10% 12.20% 11.20% Source: Rudolph, "Why Rajiv Gandhi's Death Saved the Congress," pp. 446-7.  The day after Rajiv Gandhi's murder, when the results already showed that the country would be heading towards electing yet another hung parliament, President Venkataraman proposed that a National Government be formed after the polls. Although some senior BJP and Congress(I) leaders, in particular Atal Bihari Vajpayee and Vasant Sathe respectively, received the proposal favorably the idea was quickly rejected by all Lloyd I. Rudolph, "Why Rajiv Gandhi's Death Saved the Congress: How an Event Affected the Outcome of the 1991 Election in India", in Harold A. Gould and Sumit Ganguly, eds., India Votes: Alliance Politics and Minority Governments Tenth Elections (Boulder: Westview Press, 1993), pp. 436-54. Ibid., p. 440. 166  167  in the N  83  parties. The main argument against the formation of a National Government was that this idea was based on European experience where  political parties ... operate on minimum consensus among themselves even in normal times, and in exceptional circumstances they can come together on the basis of such a consensus. Therefore, the essence of National Government lies in minimum policy and programmatic consensus among the major political parties of a country. It has been further argued that such a thing does not exist in India. The major political parties in India are competing against one another and every major political party is confronting opponents with opposite policies and programmes. The Congress(I), the Janata Dal, the National Front, the BJP and the Left political parties, with their conflicting ideologies, cannot be expected to sit together to evolve a common economic strategy for India. If the Left parties were against the IMF loan, all other centrist parties agreed on its necessity. The BJP opposed the introduction of the Places of Worship (Special provisions) Bill while all other parties supported it. 168  Concurrently with the national polls, the states of Assam, Haryana, Kerala, Tamil Nadu, Uttar Pradesh and West Bengal and the Union Territory of Pondicherry also held elections to their Vidhan Sabhas. The Congress(I) performed very well in these elections. It won majorities and went on to form governments in Assam, Haryana, Pondicherry, and Kerala.  169  In Tamil Nadu, the Congress(I)'s electoral ally, the AIADMK, formed a  majority government, whereas in Uttar Pradesh and West Bengal the Congress(I) performed very poorly. In the former state the BJP, while in the latter the CPI(M)-led Left Front, formed governments. Following the election, the President invited the leader of the Congress(I), Narasimha Rao, to form a government by virtue of his party's being the largest in the Lok Sabha. Since the Congress(I) was not in a majority position its decision whether to accept or decline the invitation was sensitive to the other parties' attitudes to the formation of a CongressfJ) government. M . L . Ahuja and Sharda Paul, 1989-91 General Elections in India (New Delhi: Associated Publishing House, 1992), pp. 127-8. In Kerala, it was the Congress(I)-led United Democratic Front that won a majority and formed a coalition government. , 6 8  169  84  The BJP pledged that it would act as an effective and responsible opposition to a Congress(I) government. The party leadership made it clear that although by being the official opposition the BJP regarded it to be its duty to vote against the Congress(I) should it attempt to form a government, it wanted no early elections and preferred the cabinet to last a full term in office.  170  It was reported that Narasimha Rao and Lai Krishna  Advani, the respective leaders of the Congress(I) and the BJP, agreed on a compromise package: while the BJP would support the Congress(I)'s economic policy initiatives in the Lok Sabha, the Rao government would in turn provide assistance to the state governments of Madhya Pradesh, Rajasthan, Hirnachal Pradesh and Uttar Pradesh each controlled by the BJP.  171  The non-BJP parties were ambivalent about the formation of a Congress(I) government. On the one hand, they also wanted government stability and no more elections after just having been through two rounds in as many years. On the other hand, however, the National and Left Fronts were also concerned that the Congress(I) might be able to develop a strong record in office and improve its reputation among the electorate.  172  Furthermore, the parties were also concerned that a Congress(I)-led national  government would make subnational governance difficult in West Bengal, where the CPI(M)-led Left Front had controlled the state government for the last 20 years, and in Bihar and Orissa, both controlled by the Janata Dal.  173  It was reported that the Congress(I)  leadership did pressure the National Front to cooperate with its government by threatening to dismiss the Bihar government marred by severe law and order problems.  174  Eventually, both the National Front and the Left Front decided to support a Congress(I) on an issue-to-issue basis.  ™India Today (July 15,  1991) p. 43. Yogendra Malik and V. B. Singh, Hindu Nationalists in India: The Rise Bharatiya Janata Party (New Delhi: Vistaar Publications, 1994) p. 91. }12 India Today (July 15, 1991), p. 15. m India Today (July 31, 1991), p. 23. 171  Ibid, p. 25.  UA  of the  85  With all these developments in the background, Narasimha Rao accepted the President's invitation to form a government. His minority administration passed the vote of confidence, as instructed by the President, on July 15 with 241 legislators voting in favor, 111 against the government and 112 abstaining. The margin by which the 175  Congress(I) fell short of a majority in the Lok Sabha improved slightly after by-elections in November 1991 and the holding of elections in Punjab in February 1992. O f the 15 Lok Sabha seats that were contested in the by-elections the Congress(I) won 8. In Punjab, with the Shiromani Akali Dal, the main opposition party to the Congress(I) in the state, boycotting the polls, the CongressfJ) won 12 of the 13 seats.  176  The CongressfJ) managed to remain in office for the entire duration of its term by constantly seeking out new alliances with different parties on each issue that had to be voted on. For example, to have its own nominee, Shivraj V. Patil, elected as Speaker of the Lok Sabha, the CongressfJ) made a deal with the BJP right after the election.  177  However, the BJP voted against the government on its investiture vote while both the National and the Left Fronts abstained. On March 9, 1992, the CongressfJ) secured to 178  have its motion of thanks to the President's speech opening the budget session of parliament passed even though neither the BJP, nor the National nor the Left Front supported it. In the face of 231 votes against the motion, the government rallied 269 votes in support of its motion, while 32 legislators abstained. Among those abstaining were 9 Telugu Desam representatives who openly declared their support for the Rao government in defiance of their party's official position.  179  Although the relationship between the CongressfJ) and the BJP quickly moved from consensual cooperation to bitter opposition, the Rao government managed to Robert L . Hardgrave, "Alliance politics and Minority Government: India at the Polls, 1989 and 1991," in Harold A. Gould and Sumit Ganguly, eds., India Votes, p. 237.  175  Ibid  116  ™India Today, July 31, 1991, p. 25. ™Keesing's Contemporary Archives (1991), p.38337. Hardgrave, "Alliance Politics and Minority Government," p. 237. 179  86  survive the rift. In order to embarrass the government, Arjun Singh, a factional opponent of Narasimha Rao in the Congress(I) Party, issued a sharp criticism of the BJP arguing that the Congress(I)'s traditional policy of secularism was at odds with the party cooperating with the BJP in the Lok Sabha. In order to preserve unity within his own 180  party, Rao echoed that communal parties have no place in a secular state, which was how the Constitution defined India, and went even further by calling for a legal ban on nonsecular parties.  181  However, by so doing he also encouraged the hard-liners within the  BJP, led by Murli Manohar Joshi, who had criticized the party's moderate wing for its cooperation with the Congress(I). At its May 1992 meeting in Gandhinagar, the BJP National Council reaffirmed the party's commitment to Hindu nationalism in general and the construction of the temple at Ayodhya in particular. On December 6, 1992 thousands of volunteers mobilized by the BJP, the RSS and the Vishwa Hindu Parishad (World Hindu Council) demolished the Babri Masjid, the mosque, precipitating communal riots all over the country. In response, the Congress(I) government banned the RSS and the VSP, authorized the arrest of over 5000 BJP party officials, and dismissed the four state governments controlled by the B J P .  182  When the  BJP sponsored a no-confidence motion against the government, the Left voted with the Congress(I) while the National Front helped the government by abstaining.  183  The alienation of the BJP left the Congress(I) in a precarious position. As long as the issue of secularism versus communalism did not surface in the Lok Sabha, the government could count on the BJP's support on economic policy issues. However, the Ayodhya incident broke this arrangement leaving the government vulnerable to attacks from both the right and the left. Nonetheless, the Congress(I) managed to remain in office  Singh waged his attack at the famous Tirupathi session of the All India Congress Committee in April 1992. See, Malik and Singh, Hindu Nationalists in India, p. 92. 180  Ibid  m  Malim and Singh, Hindu Nationalists in India, p. 95. * Keesing's Contemporary Archives, 38 (1992), p.39222.  182  x  3  87  by taking advantage of the fragmentation of the party system and securing the ad hoc support of small splinter groups from time to time as political expedience required. Thus, the Congress(I) survived the no-confidence motion sponsored by the BJP as well as both the National and Left Fronts in July 1993 as the 3 member Samajwadi Party (Mulayam) and the 6 member BSP abstained while 4 members of the Jharkhand Mukti Morcha and 7 members of the Janata Dal (A) voted with the government against the motion.  2.8  184  T H E E L E V E N T H L O K SABHA, 1996-97: T H R E E U N S T A B L E MINORITY GOVERNMENTS Similarly to the 1991 polls, the 1996 election was a triangular fight among the  National Front-Left Front, the Congress(I)-, and the BJP-led alliances. By winning a mere 26.4% of the seats and 28.8% of the popular vote, the Congress(I) showed its worst ever electoral performance since Independence. While it still remained the largest party in terms of its share of the popular vote, it was only the second largest party in terms of its share of the Lok Sabha seats after the BJP, which won 34.2% of the seats but only 20.3% of the votes.  185  Shortly before the elections, a number of influential leaders had left the Congress(I) over their disagreements with Narasimha Rao and formed new political parties: the Congress (Tiwari), the Karnataka Congress, the Madhya Pradesh Vikas Congress and the Tamil Maanila Congress.  186  Whereas the Tamil Maanila Congress  India Today, August 15, 1993, pp. 38-42. Allegations that the Rao government literally bought the support of the J M M on this vote were confirmed in police investigations in 1996. These investigations precipitated the resignation of Rao from the Congress(I) leadership. XM  Statistical Report on General Elections, 1996 to the Eleventh Lok Sabha (New Delhi:  185  Election Commission of India), Table LS-21. The Congress(Tiwari) was formed by Arjun Singh and N . D. Tiwari who had criticized Rao's leadership for abandoning the party's traditional leftist and secular policy position. The Karnataka Congress was formed by former Chief Minister Bangarappa who left the party because Rao supported his factional enemies in the state party leadership. The Tamil Maanila Congress was formed in March 1995 by a faction, led by G. K. Moopanar, 186  88  entered into an electoral agreement with the National Front in the state of Tamil Nadu, the other parties did not ally with either the BJP or the National Front-Left Front.  187  Concurrently with the national election, State Assembly elections were also being held in the states of Assam, Haryana, Kerala, Tamil Nadu and West Bengal, and in the Union Territory of Pondicherry. In these state elections, the Congress, that had controlled each of the outgoing state governments except for that of West Bengal, fared miserably. In Haryana, the B JP-Haryana Vikas Party alliance won, while in the rest of the states various members of the National Front-Left Front triumphed. In Assam they were the Asom Gona Parishad in coalition with the Autonomous State Demand Committee, the CPI, the CPI(M) and the UPPA; in West Bengal and Kerala it was the Left Front, and in Tamil Nadu and Pondicherry they were the D M K in coalition with the T M C and the CPI. In Tamil Nadu the coalition also included the INL and the All India Forward Bloc (Ayyannan Ambalam). The BJP formed electoral alliances with the Shiv Sena, the Samata Party  188  and  the Haryana Vikas Party. These alliances, however, were quite contradictory. In the state of Haryana, the BJP entered into an electoral agreement with the Haryana Vikas Party both for the Lok Sabha and the Vidhan Sabha polls. However they both ran candidates against candidates fielded by the Samata Party in each of the ten constituencies of the state. In six of the ten constituencies the BJP ran against a Samata candidate and in four  in the state unit of the Congress(I) in Tamil Nadu after Rao decided, against the state party's warning, that the Congress(I) would contest the 1996 election in alliance with the AIADMK. The Madhya Pradesh Vikas Congress was formed by the prince of Gwalior, Madhav Rao Scindia, whose mother, Vijayaraje Scindia, is one of the Vice-Presidents of the BJP. Scindia left the party because he was alleged to have been involved in a foreign exchange scandal as a result of which Rao would not have given him a Congress(I) ticket to run on. The only exception is the Madhya Pradesh Vikas Congress that fielded only one candidate, the party's leader Madhav Rao Scindia, in whose favor both the BJP and the National Front-Left Front withdrew. Meenu Roy, India Votes ... p. 85. The Samajwadi Janata Party, that had formed a short-lived national government in 1990, merged into the Samata Party in the spring of 1996 shortly before the elections. 187  188  89  constituencies the Haryana Vikas conflicted with the Samata. In contrast, the BJP arrived at a perfect coordination of its candidate nomination with the Samata Party in Bihar where the former fielded 32 and the latter 22 candidates. Although the two parties did not run a joint campaign in Andhra Pradesh they did not field candidates against one another either. However, in every other state where the Samata entered the fray, these states being Gujarat, Himachal Pradesh, Kerala, Madhya Pradesh, Maharashtra, Manipur, Orissa, Rajasthan, Tamil Nadu, Uttar Pradesh, West Bengal, Delhi and the Union Territory of Chandigarh, the two parties did nominate candidates running against each other's. The BJP and Shiv Sena, similarly to the BJP-Haryana Vikas Party alliance arrived at perfect coordination in the state of Maharashtra. The BJP nominated 24 and the Shiv Sena 20 candidates, none in the same constituency. The BJP failed to renew its traditional alliance with the Shiromani Akali Dal in the Punjab although the two parties had had a long history of both electoral and legislative cooperation in that state. Nevertheless, although the two parties ran candidates against each other in 3 of the 12 constituencies in Punjab, the Akali Dal offered its support from the outside to a BJP government, if one could be formed, in the Lok Sabha. The National and Left Fronts formed the United Front after the elections even though their various members did not cooperate with one another consistently in the electoral stage. In particular, while the principal parties of both Fronts, the Janata Dal and the CPI(M) respectively, sought to continue their partnership, they ran into a number of conflicts during their effort to build and expand the joint alliance. First, the National Front was not interested in cooperating with the Samajwadi Party. Ever since the latter caused the collapse of the incumbent National Front government in the Lok Sabha in 1990 and went on to form a government relying on the support of the CongressfJ), the relationship between the two parties had been sour. Therefore the Janata Dal was in favor of forming an electoral alliance with the Bahujan Samaj Party, which was the main opponent of the Samajwadi Party in the state of Uttar  90  Pradesh. However, the Left Front found cooperation with the Bahujan Samaj totally 189  unacceptable because the party had relied on the support of the BJP, the arch enemy of the Left, to form a government the year before. Ultimately, the Left Front prevailed and the National Front agreed to engage in electoral coordination with the Samajwadi Party. Second, the question of admitting the Akali Dal (Badal) into the alliance also caused friction between the National and Left Fronts. Whereas the Left wanted to have no truck with a communal party,  191  the National Front was aware that in order to defeat  the Congress(I) in Punjab, the formation of an electoral alliance with the Akali Dal was imperative. However, once again the Left prevailed and the Akali Dal (Badal) entered into an alliance with the BSP. The two parties virtually swept the polls in Punjab by winning 11 of the 13 seats contested there.  192  The third source of friction between the National and the Left Fronts was caused by the split in the Telugu Desam, one of the founding parties of the National Front, whose ex-leader, N . T. Rama Rao had been the Front's convener. In August 1995 Chandrababu Naidu, Rama Rao's son in-law masterminded the replacement of Rao as leader of the Telugu Desam legislature party in the Andhra Pradesh State Assembly. As a result, the party split into two rival factions, one led by Naidu and the other by Rao, and following his death in January 1996, his widow Laxmi Parvathi. In the national polls, the National and Left Fronts ended up forming different alliances with the splinter parties of the Telugu Desam: the left cooperated with Naidu, whereas the National Front Although the Bahujan Samaj Party and the Samajwadi Party had contested the state election of Uttar Pradesh in alliance in December 1993, and subsequently formed a coalition government headed by Mulayam Singh Yadav, their cooperation broke down in the summer of 1995 when the Bahujan Samaj Party withdrew from the coalition causing its collapse. In the aftermath, the Bahujan Samaj Party was assisted by the BJP from the outside to form a minority government in the Uttar Pradesh Vidhan Sabha. Meenu Roy, India Votes: Elections 1996, A Critical Analysis (New Delhi: Deep and Deep Publications, 1996), p. 38 and 79-84. 189  190  The avowed purpose of the Akali Dal is to represent the aspirations of the Sikhs in the state of Punjab. Roy, India Votes: Elections 1996, pp. 38, 72-6, and 198-9.  ,91  192  190  91  cooperated with Parvathi. Of the two alliances the former proved to be more successful, however, the inability of the two Fronts to overcome their divisions contributed to the CongressfJ) relatively strong electoral performance in the state. Of the 41 seats contested, the CongressfJ) won 22, the Telugu Desam (Naidu) bagged 16, while the CPI, the CPIfM) and MTM won 1 seat each.  193  Fourth, in the state of Tamil Nadu, the National and Left Fronts were divided not only between but also within themselves. As for the Left, two of its constituents, the CPI and the Forward Bloc, entered into an agreement with the Dravida Munnetra Kazagham, which had already had an understanding with the Janata Dal, the Tamil Maanila Congress, and the Indian National League. In contrast, the CPI(M) formed an alliance with the Marumalarchi Dravida Munnetra Kazagham. The Congress (Tiwari), which eventually, came to enter the United Front, allied with the Pattali Makkal Katchi (PMK), a splinter party from the D M K . Lastly, in Maharashtra state, the National Front-Left Front combine formed two separate alliances. Whereas the Janata Dal entered into electoral coordination with the Peasants' and Workers' Party and splinter factions of the Republican Party of India, the Left Front allied with the Samajwadi Party and a new party born of the merger of a faction of the Republican Party of India and the Bahujan Maha Sangh.  194  As the BJP emerged from the election as the single largest party in the legislature, the President invited it to form a government on May 15. Throughout the campaign the BJP had indicated that it would not seek to form a government unless it would win at least 220-225 seats in the Lok Sabha. Thus, it came as quite a surprise when the BJP 195  accepted the President's invitation even though it won only 161 seats. On May 16 Atal Bihari Vajpayee, the BJP's parliamentary leader, was sworn in as Prime Minister. His  India Votes, p. 114-8 and 229. ™Ibid, p.93. India Today (June 15, 1996), p. 40. 193  i95  Roy,  92  cabinet ministers including one member of the Shiv Sena were also sworn in at the same time. The Shiromani Akali Dal, the Haryana Vikas Party and the Samata Party offered their support to the government without accepting any ministerial portfolios. President Shankar Dayal Sharma instructed Vajpayee to prove his cabinet's majority by May 31.  196  Vajpayee claimed that the party decided to proceed with forming a government in order t o keep the Third Front, which by then had been renamed the United Front, and the CongressfJ) polarized. In particular, the BJP hoped that it would be able to secure the support of parties like the D M K , the T M C , the TDP, and AGP by appealing to their subnational record of opposition to the CongressfJ) as well as offering them a chance t o share power at the national level by participating in the BJP-led cabinet.  197  The BJP also  sought to gain the cooperation of the CongressfJ): in exchange for the CongressfJ) abstaining from voting on the confidence motion the BJP was ready to support the reelection of Shivraj Patil of the CongressfJ) as Speaker of the Lok Sabha. In addition, 198  Vajpayee also claimed that  when the president invited me [Vajpayee] to form a government, the political situation was fluid. The regional parties were keeping their options open. We [the BJP] wanted to make an honest effort to form the government in the light of the people's mandate, with the help of regional parties on the basis of a common minimum programme. 199  The fluidity of the political scene, however, started to thicken as the National Front, the Left Front, and the four ex-Congress(I) splinter parties formed the United Front which tabled a no-confidence motion against the BJP government on May 22. The debates on the motion took place on May 27 and 28. As not a single party seemed to change sides during the debate, it became apparent that the BJP government would not survive the vote. Before the motion would have been put to a vote, Vajpayee tendered the  India Votes, p. 250. India Today (June 15, 1996), p. 41. *India Today (May 31, 1996), pp. 31-33, and (June '"India Today (June 15, 1996), p. 41. 196  Roy,  191 19  15, 1996), pp. 40-41.  93 resignation of his government to the President after only 13 days in office.  200  Following the resignation of the BJP cabinet and the expressed disinterest of the Congress(F) in forming a government, the President invited H . D. Dewe Gowda, the leader of the United Front, to form a cabinet. Gowda and his cabinet were sworn in on June 1 and with the support of the Congress(I) the new government won the vote of confidence on June 12.  201  Although the Congress(I) and the two largest constituents of the United Front, the Janata Dal and the CPI(M), ran a campaign heavily attacking each other their policy programs also contained some overlaps. Thus, on the one hand the Janata Dal accused the Congress(I) of being " fossilized by family rule and the profligacy of power", of being involved in a number of scams connected to the disinvestment of public sector enterprises, of having bribed minor parties to sustain the previous minority Congress(I) government in office by voting for it or abstaining on crucial votes of confidence, and of creating communal tension between the Muslim and Hindu communities byfirstallowing the demolition of the Babri Masjid by Hindu militants and then not meting out the appropriate penalties on them.  202  Similarly, the CPI(M) also dismissed the record of the  outgoing administration by calling it a "bonanza for big business and misery for the masses" and "unabashed succumbing to the pressures of US imperialism." The 203  Congress(I) accused the Janata Dal of triggering caste conflicts, particularly in the state of Bihar, and the entire opposition of propagating the idea of an inward-looking premodern India, and failing to provide stable national governments in the past.  204  On the other hand, both sides agreed that economic growth had to be slowed down for the time being in order to put into effect programs and policies that would  India Votes, p. 253. India Votes, pp. 253-4. ™India Today (June 15, 1996), p. 28. Ibid., p. 28. Ibid., p. 28. Roy, ^Roy, 200  203 204  94  redistribute a share of the wealth to the poor masses. In addition, both the United Front and the Congress(I) agreed that the distribution of resources available to the central and the state governments should be improved and be made more equitable, and that minorities' rights should be accorded greater attention and priority. Finally, the United Front wholeheartedly subscribed to the idea put forward by Manmohan Singh, the Finance Minister in the outgoing Congress(I) government, that the small-scale sector of the economy must be strengthened and relied on for economic growth, employment creation and import-substitution.  205  The United Front parties established a Steering Committee that was in charge of coordinating the legislative strategies of the Front members. Parties that declined to enter the government formally, most conspicuously the CPI(M), were also given representation and say in the Committee's decision-making. In addition, the Front and the Congress(I) 206  also created a coordination committee in order to develop and pursue a common agenda, however, it was agreed by both parties that their respective leaders, Dewe Gowda on the one hand and outgoing Prime Minister Narasimha Rao on the other, should have a great deal of flexibility in shaping the relationship. The cabinet that was sworn in on June 1 included ministers from only 9 parties: the Janata Dal, the Samajwadi Party, the D M K , the TDP, the AGP, the INC(T), the M P V C and the KCP. Similarly to the Congress(I), the four parties of the Left Front as well as the IUML, the J M M , the MTM, the MGP, the UGDP, the A S D C , the SDF, the K E C and 3 Independents provided external support to the coalition. On June 28 the cabinet was expanded and the other members of the United Front, with the exception of the CPI(M), also took up ministerial portfolios.  207  India Today (June 15, 1996), p. 25. ^Roy, India Votes, p. 254; India Today (June 15, 1996), p. 25. J . C. Aggarwal and N. K. Chowdhry, Elections in India 1952-1996: Constituency Profiles, Results and Analysis Focussing Poll 1996 (Delhi: Shipra Publications, 1996), p m  207  287-8.  95  Soon after the United Front took office in May 1996 the CongressfJ) underwent a change in its top leadership. The President of the Party, Narasimha Rao, resigned from his post under increasing pressure from the CongressfJ) Working Committee which blamed Rao for alienating important CongressfJ) politicians that in turn had contributed to the CongressfJ)'s debacle in the 1996 parliamentary elections. His successor was the long-time treasurer of the party, Sitaram Kesri. Amongst the first measures that Kesri took was to lure back the party's rebels that now belonged to the United Front. The first to return was the Karnataka Congress Party, followed by the Madhya Pradesh Vikas Congress and the Indian National Congress (Tiwari). Although the Tamil Maanila Congress was also intensely lobbied it did stay with the United Front. In defense against the Congress(I)'s maneuvers, Prime Minister Deve Gowda sought to maintain his government's autonomy and refused to yield to the various demands posed by the CongressfJ), which expected to be consulted on government policy in reciprocation for the legislative support it was extending to the cabinet. There were two particularly important cases in point that severed the relationship between the United Front and the Congress(I). The first one was the Vidhan Sabha election in the state of Punjab.  208  This election was particularly important for Kesri as it  offered him the opportunity to demonstrate his ability to lead the party to electoral success. The Congress(I) demanded that the United Front help it by forming an electoral alliance against the Shiromani Akali Dal, which had boycotted the last state election in 1992, which entered into an alliance with the BJP. However, the United Front refused to do so and the election resulted in a humiliating defeat for the Congress(I) as its seat share dropped from 74.4% to 11.9% in the new Assembly. Although Punjab was not an important state for the United Front, it was for the  Hndia Today (February 28,  1997), pp. 56-61.  96  Congress(I), which was in power in only two major states: Orissa and Madhya Pradesh. The non-compliance of the United Front with the Congress(I)'s request also added to the growing tension within the Congress(I) Party. The local Congress(I) units in those states where a United Front constituent was in power, namely Andhra Pradesh, Assam, Bihar, Haryana, Karnataka, Kerala, Tamil Nadu, West Bengal, felt increasingly uneasy about their party's national wing maintaining their opponents in power at the national level.  209  The second clash between the Front and the Congress(I) took place in Uttar Pradesh, where the state election held in September 1996 had resulted in a hung assembly. The Congress(I) had contested the election as the junior partner in alliance with the Bahujan Samaj Party, the party whose withdrawal had caused the collapse of the coalition government dominated by Mulayam Singh Yadav's Samajwadi Party the year before. The Congress(I) demanded that the United Front support the formation of a Bahujan Samaj-Congress(I) coalition government with the former providing the Chief Minister. This proposal, however, was unacceptable to the United Front because of the antagonism between the Samajwadi Party, its principal constituent in Uttar Pradesh, and the Bahujan Samaj Party.  210  As no combination of parties managed to put together a  working majority, the central government imposed President's Rule on Uttar Pradesh. On March 19, after six months of incessant negotiations, the Bahujan Samaj Party and the BJP agreed to form a coalition cabinet with the Chief Ministership rotating between the two partners. No sooner did the news break, the Congress(I) started 211  courting the Tamil Maanila Congress in an effort to topple the United Front government. Acting as Kesri's personal envoy, senior Congress(I) leader K. Karunakaran, a former Chief Minister of Kerala state, was reported to have offered G. K. Moopanar, the leader of the Tamil Maanila Congress, the Prime Ministership should it form a coalition with the  ^Frontline (April 18, 1997), p. 5 See footnote 153. ^Frontline (May 2, 1997), p. 15. 210  97  CongressfJ). On March 29, Karunakaran told a press conference that Moopanar's response to the offer was positive.  212  The next day, Sitaram Kesri notified President  Sharma that the CongressfJ) had decided to withdraw its support from the United Front cabinet in the Lok Sabha. In turn, President Sharma instructed Deve Gowda to hold a vote of confidence by April 7.  213  The confidence motion, which was put to a vote on April  11, was defeated by 190 in favor and 338 against as the CongressfJ) joined the BJP in voting against it. Consequently, Prime Minister Gowda submitted his and his cabinet's resignation to the President. As soon as the Gowda cabinet was instructed by the President to seek a vote of confidence in the Lok Sabha, the BJP floated the idea of extending its alliance and forming a National Democratic Front including more regional parties in an effort to make another attempt at forming a government.  214  At the meeting of the party's national  executive, a few days before the confidence vote, Lai Krishna Advani openly identified the Congress, the Communist parties, the Samajwadi Party and the Janata Dal as the BJP's adversaries with whom the party could not form an alliance of any sort. In an effort to court the support of the regional and other minor parties, Advani called on his party to "debunk the propaganda that the BJP was an anti-Dalit and anti-minority party." He 215  further called on his party to identify more strongly with regional sentiments and concerns without consenting to regional chauvinism and to strengthen the party's ties with its present regional allies. The BJP's claim to be more sensitive to the interests of minority and regional parties was made credible by the party's successful cooperation with the Shiromani Akali Dal and the Bahujan Samaj Party to form coalition governments in the Punjab and Uttar Pradesh respectively. In any event, there was no significant response from the other  ^Frontline (May 2, 1997), pp. 12-3; Frontline (April ^Frontline (April 18, 1997), p. 5. ^Frontline (May 2, 1997), pp. 18-9. Ibid, p. 19.  21$  18, 1997), p. 11.  98  regional parties in the United Front to the BJP's proposal to form a broader front and thus the BJP leadership decided not to stake an official claim to form a government. The Congress(I) was also interested in exploring the option of forming a government. At first, the party demanded that the United Front assist it to form a government with the party's president Sitaram Kesri becoming Prime Minister. As the United Front did not agree, the Congress(I) proposed the formation of a Congress(I)United Front coalition in which the two partners would have an equal standing, however, the United Front could provide the Prime Minister. Finally, upon the leaking of two letters sent by the United Front to the President indicating the Front's determination to prevent both the BJP and the Congress(I) from entering office, the Congress(I) Working Committee decided on April 14, 1997 to renew its support to the Front rather than risk new elections. The only conditions that the Congress(I) posed was that Gowda had to be replaced as Prime Minister first.  216  The Congress(I)'s preferred candidate for the United Front leadership was G. K . Moopanar, leader of the T M C . The Congress(I) would have liked to see Moopanar leading the United Front not only because of his strong links with his old party but also because Moopanar had made hints in the past at allowing the Congress(I) to share power with the United Front at the right time. However, exactly for this reason, Moopanar was 217  not acceptable to the other members of the United Front who shared a strong antiCongress position in common. Furthermore, M . Karunanidhi, the leader of the D M K and Chief Minister of the state of Tamil Nadu was also concerned that if Moopanar were Prime Minister, the balance of powers between T M C and the D M K , who formed coalition government in Tamil Nadu, might be upset. In the end, I. K . Gujral, a senior member of the Janata Dal and Minister of External Affairs in the Gowda cabinet, was chosen to follow Gowda as the leader of the ^Frontline (May 2, 1997), pp. 4-9. Ibid., p. 12. 217  99  United Front. The Congress(I) accepted him and a new United Front cabinet was sworn in and won a vote of confidence on April 21. The reformed United Front cabinet with I. K. Gujral at its helm did not last long in office. While the government survived the breaking away of a faction led by Laloo Prasad Yadav and the embarrassment that President Narayanan's refusal to impose 218  President's Rule on Uttar Pradesh caused , it collapsed after the publication of leaked 219  excerptsfromthe report of the Jain commission, investigating the assassination of former Congress(I) Prime Minister Rajiv Gandhi, led to the Congress(I)'s withdrawal of support once again. According to the leaked information, the Jain commission accused the DMK for maintaining connections with and providing support to the Sri Lankan rebel group, the Liberation Tigers of Tamil Eelam (LTTE) that had plotted the assassination. In immediate response to the publication of these excerpts the Congress(I) demanded on November 20 that it would withdraw its support from the United Front  0n July 3 Laloo Prasad Yadav, the national president of the Janata Dal and Chief Minister of the state of Bihar was defeated in the party's national leadership contest by Sharad Yadav who won 691 votes against Laloo's 58. Protesting that the voting was rigged Laloo Yadav rejected the result, broke ranks with the Janata Dal andfloateda new party called the Rashtriya Janata Dal. The formation of this new party did not pose an immediate threat to the United Front coalition cabinet because Laloo could only muster the support of 16 Janata Dal legislators in the Lok Sabha, all from his home state of Bihar, and 8 members of the Rajya Sabha. Later on at the end of the month, Laloo was arrested on charges of his involvement in an animal fodder scam in Bihar. Upon his arrest, Laloo appointed his wife, Rabri Devi as Chief Minister of Bihar, who was later confirmed in her post by a vote of confidence passed by the Bihar Vidhan Sabha. In October 1997 when President K. R. Narayanan rejected the Prime Minister's advice to impose direct President's Rule on the state of Uttar Pradesh. Gujral sought the imposition of President's Rule on the state in the aftermath of the collapse of the Bahujan Samaj Party-BJP coalition government there and the popular violence it had sparked. Although the Bahujan Samaj Party had withdrawn from the coalition on October 19, barely a month after it had transferred the rotating Chief Ministership to the BJP on September 21, in the subsequent vote of confidence on October 21, the BJP cabinet survived as the opposition parties, prominently the Bahujan Samaj and the Congress(I), boycotted the session. President Narayanan's refusal to heed the Prime Minister's advice forced the cabinet to reconsider its position and on October 22 it resolved no to pursue the imposition of President's Rule on the troubled state. 218  219  100  unless the D M K were expelled from Front and excluded from the cabinet. The Congress(I) claimed that it could not justify supporting a government a constituent party of which had been directly responsible for the assassination of the Congress(I)'s former leader. After Gujral refused to bow to the Congress(I)'s demand on November 24, Sitaram Kesri notified the President of his party's terminating its support of the United Front cabinet as of November 28.  220  The decision of the Congress(I) leadership infuriated the party's MPs, many of whom were novices inthe Lok Sabha and who were concerned that new elections might be called in resolution of the crisis. The MPs resented that after having incurred huge electoral expenses only 20 months ago now they were risking the loss of their seats by having to go through the costly electoral exercise once again. A number of Congress MPs demanded that the party automatically renominate them if elections were called and that the official list of candidates be published early so that there would be sufficient time to campaign.  221  Sensing the resentment of many Congress(I) MPs, the BJP tried hard to win their support over and experiment with the formation of a new BJP cabinet in the Lok Sabha. On December 3, Lai Krishna Advani had told a press conference that his party has had the support of 40 Congress MPs who were ready to break ranks with their party. Although, as Advani noted, it had never happened before that so many Congress(I) MPs would express their dissatisfaction with their party's leadership and their readiness to join another political party in the Lok Sabha, this number was not sufficient to avoid the penalties under the Tenth Schedule. Thus, the BJP resolved that if it could not get Congress(I) representatives to change sides, it would recommend to the President that the Lok Sabha be dissolved and new elections be held.  ^Keesing's  Contemporary Archives (1997), p. 41914.  TheHindustan Ibid  m  m  222  Times, www2.hindustantimes.com/nonfram/041297/detFR003.htm  101  Congress(I) leaders who were adamant against the holding of new elections tried to work out a compromise formula with the United Front. In particular, Sharad Pawar, the leader of the Congress(I) Legislature Party met with A. B. Bardhan of the Communist Party of India suggesting that if the Dravida Munnetra Kazagham withdrew from the coalition briefly until it was cleared of the Jain Commission's indictment then a rapprochement between the Front and the Congress(I) could be possible. However, no agreement on the formula could be achieved.  223  In the evening of December 3, the cabinet decided to advise the President to dissolve the Lok Sabha and order fresh elections. The President concurred and asked I. K. Gujral to remain in office in caretaker capacity until the new government was formed.  The Hindustan Times, www2.hindustantimes.com/nonfram/041297/detFR004.htm  102  CHAPTER THREE: PARTY SYSTEM CHANGE The Indian National Congress Party dominated the national party system of Independent India from 1952 until 1989. With the sole exception of 1977, the party won legislative majorities in every general election in this 37 year period. Although the Congress Party was defeated in 1977 it quickly regained its dominant position in the subsequent election held in 1980. In contrast, since 1989 the Congress Party's share of the popular vote has shown a secular decline and the party has repeatedly failed to win a majority of the seats in the Lok Sabha. The decline of the dominant party has been accompanied by the inability of any other party to win a legislative majority thus ushering in a period of successive hung parliaments in India. This change in the pattern of election outcomes has led scholars to herald the arrival of genuine multipartism in India's national party system.  224  This chapter seeks to explain the transformation of the Indian party system from predominance to competitive multipartism. It will argue that whereas the main Indologist and comparative theories of party system change fail to provide an adequate and consistent account for changes in the format of the party system, an extension of William Riker's theory of sophisticated and sincere voting under the single-member-simple plurality electoral formula does provide such an explanation. In particular, the chapter 225  will demonstrate that both the recovery and the decline of the Congress after the 1977 and the 1989 elections respectively are the result of such voting behaviour. Furthermore, the chapter will also show the importance of institutions and elites: the former determined the actual form of electoral cooperation that the opposition elites adopted, merger in 1977 and alliance in 1989, and the latter in turn created the context for both sophisticated and  ^"Mahendra Prasad Singh, "From Predominance to Polarized Pluralism: the Indian Party System", Think India 2 (1990), pp. 70-8; Csaba Nikolenyi, "The New Indian Party System: What Kind of a Model?" Party Politics 4 (1998), pp. 367-80. ^William H. Riker, "The Number of Political parties: A Reexamination of Duverger's Law", Comparative Politics 9 (1976), pp. 93-106.  103  sincere voter choice. 3.1  CONCEPTUALIZING T H E D E P E N D E N T V A R I A B L E : P A R T Y S Y S T E M CHANGE Conventional wisdom in the Indologist literature has identified six stages in the  evolution of the Indian party system: 1) 1947-1967: the era of one-party dominance by the Congress Party, also referred to as the Congress system; 2) 1967-71: the first coalitional period; 3) 1971-1977: the second Congress-dominance period; 4) 1977-1980: second coalitional period; 5) 1980-1989: third Congress-dominance period; 6) 1989 onwards: multipartism.  226  The first period was characterized by the Congress Party's winning majorities in both the national and subnational elections, as shown in Figure 3.1. Although the Congress never received a majority of the popular vote cast it managed to benefit from the seat bonus that the single-member simple plurality electoral system conferred on it as a large party with solid nation-wide following.  227  The election of 1967 marked the  beginning of the first coalition period.  ^This periodization was offered by Ram Joshi and Kirtidev Desai, "Towards a More Competitive Party System in India", Asian Survey 18 (1978), pp. 1091-116. It has been brought up to date in M . P. Singh, "From Centrist Predominance to Polarized Pluralism: the Post-1989 Party System in India", in M . P. Singh and Rekha Saxena, eds., India's Political Agenda: Perspectives on the Party System (Delhi: Kalinga Publications, 1996), p. 35; and Rekha Saxena, "Party System in Transition", in M . P. Singh and R. Saxena, eds., India's Political Agenda... pp. 49-81. ^The single member simple plurality system tends to reward disproportionately large parties with solid support all across the country and regionally concentrated small parties. The logic of this is quite simple. According to this electoral formula, in each electoral district the candidate with the largest number of votes wins the single seat that is being contested. Since a candidate does not need to have majority support in the district in order to win, a party whose candidates may win in a majority of the districts may thus end up securing less than a majority of the popular vote across all those districts. In this situation, the party will receive a seat bonus, i.e. its share of the parliamentary seats will exceed its proportional share of the vote. Regionally concentrated parties also tend to receive a seat bonus from this electoral formula because by winning most of the seats in the region of their influence their share of the total legislative seats will be bound to exceed their share of the total popular vote.  104  F I G U R E 3.1: T H E C O N G R E S S P A R T Y IN T H E I N D I A N P A R T Y S Y S T E M ,  1952-96  80r 60  ^  4020 1952 1957 1962 1967 1971 1977 1980 1984 1989 1991 1996 Election year % Congress vote  *  % Congress seats  Source: Butler, Lahiri and Roy, India Decides... p. 67. A s described in the previous chapter, the Congress failed to win majorities in a number o f state elections, that were held concurrently with the national polls, and it also lost its majority in the L o k Sabha. The 1971 election brought the Congress (Ruling) back to dominance, however, it gave way to an fragile and unstable two-party system in 1977 when the Janata Party formed an alternative majority government for the first time. Once again, however, the Congress, this time named the Congress(I), reasserted its dominance in the 1980 election. This period of the third Congress dominance lasted until 1989, the first general election that would produce a hung parliament. Since 1989, the Indian party system has been acknowledged to have moved to the stage of genuine multipartism. The application of comparative typologies to identifying changes in the format o f Indian party system yields a similar periodization. The single most influential such typology has been offered by Giovanni Sartori who claims that the two most important dimensions according to which party systems can be distinguished are the number o f parties in the system and the degree o f its ideological polarization.  228  Although Sartori  ^ G i o v a n n i Sartori, Parties and Party Systems: A Frameworkfor Analysis (Cambridge: Cambridge University Press, 1976), pp. 288-9.  105  treats polarization as an independent dimension of analysis, he also maintains that it is strongly correlated with the number of parties. Accordingly, he argues that most party systems with many parties will be more polarized than party systems that are characterized by the presence of a smaller number of parties. O f the various types of party systems that his framework identifies, Sartori considers the predominant party system, the two-party systems, the moderate pluralism and the polarized pluralism types to be the most common ones. FIGURE 3.2: SARTORI'S T Y P O L O G Y OF POLITICAL P A R T Y SYSTEMS Number of relevant parties 1 2 3-5  Low polarization Predominant Tw-partism Moderate pluralism  >5  Segmented multi-partism  High polarization NA Two-party polarized Limited but polarized pluralism Polarized pluralism  Note: NA means not applicable. Source: Sartori, Parties and Party Systems, 288-9.  For the present purposes, the numerical classification of party systems is more important than the ideological one. To count the number of parties in a system, Sartori proposes his famous counting rule based on the relevance of a political party. A party is relevant, according to Sartori, when it has either a coalition or a blackmail potential. With regard to the former, a party has coalition potential "no matter how small it is, if it finds itself in a position to determine over time, and at some point in time, at least one of the possible governmental majorities." As for blackmail potential, Sartori's definition is 229  somewhat less precise. He claims that a party has blackmail potential when "its existence, or appearance, affects the tactics of party competition and particularly when it alters the direction of the competition - by determining a switch from centripetal to centrifugal competition either leftward, rightward, or in both directions - of the governing-oriented parties."  230  'Sartori, Parties and Party Systems, p. 122. 'Ibid, p. 123.  106  Parties with coalition potential are relatively easy to identify: any party that is a member of a governmental majority has to be counted as such.  231  In the case of single-  party majority governments only the governing party has coalition potential. In cases of single-party minority governments, the number of parties with coalition potential includes all those parties whose legislative support is vital for the survival of the government in addition to the governing party itself. According to the same logic, when a majority-coalition government is formed, all coalition partners have to be considered relevant. On the other hand, when the government is a minority-coalition one, the coalition partners as well as the parties extending legislative support but not formally joining the government have to be counted. With regard to blackmail potential, Sartori clearly uses this criterion in order to allow for counting all parties that are large enough to matter but that have not been part of the governing coalition, either from the inside or the outside. Ware does away with the vagueness of this concept and proposes that any party that has at least 3% of the seats in the legislature should be considered relevant. In sum, then it is useful to think about relevance in terms of Ware's 3% rule of thumb coupled with the notion of coalition potential.  232  In other words, parties with more than 3% of the seats must be considered  relevant as long as they are part of a governing coalition. The concept of relevance assumes real significance only in multiparty systems. Therefore, a predominant party system is defined as such not because there is just one party that meets the aforementioned criteria but because there is one party that controls the legislature over time. Similarly, there may be more than just two relevant parties in a two-party system, which for Sartori, is defined as the alternation in office by two parties. Multiparty systems with 3 to 5 relevant parties are considered to be "limited pluralism"  Nikolenyi, "The New Indian Party System," p. 369. Alan Ware, Political Parties and Party Systems (Oxford: Oxford University Press, 1996), pp. 158-9.  231  232  107  while those with more than 5 parties are the "extreme pluralism" party systems. Figure 3.3 plots the evolution of the Indian party system according to Sartori's model. FIGURE 3.3: T H E E V O L U T I O N OF T H E INDIAN P A R T Y S Y S T E M IN TERMS OF SARTORI'S MOIDEL  Period  System type Predominant party system (The Congress Party formed every government in this period) Two-party system (The Janata Party formed a majority government in 1977) Predominant party system (The Congress(I) reasserted its dominance and formed both governments in this period) Extremely plural party system (The number of relevant parties in this period is about 25)  1952-1977  1977-1980 1980-1989  1989-present  Sartori's counting rule is not the only one that has been offered in the literature. For instance, Douglas Rae' famous fragmentation index effective number of parties  234  233  or Laakso and Taagapera's  are measures that have been widely used by scholars. The  former is an index that calculates fragmentation by deducting the sum of each party's squared vote or seat share, expressed in decimal format, from unity:  1-Spi.t  2  where pi,t is the percent of the vote or seat obtained by party i in election t. The effective number of parties index calculates fragmentation by dividing the sum of each party's squared vote or seat share, expressed in decimal format, into unity:  1-Spi.t  2  ^Douglas Rae, The Political Consequences of Electoral Laws (New haven: Yale University Press, 1971). ° M . Laakso and R. Taagapera, "Effective Number of Parties: A Measure with Application to West Europe," Comparative Political Studies 12 (1979), pp. 3-27. 4  108  Figure 3.4 plots the evolution of the Indian party system in terms of these two measures. FIGURE 3.4: T H E F R A G M E N T A T I O N A N D T H E E F F E C T I V E N U M B E R OF PARTIES IN T H E INDIAN P A R T Y S Y S T E M , 1952-96  fj  "  19E  _j;  (  "  "  1967 1962 1967 1971  1977  1980 19&  1989 1991 1996  EtecticnycH"  BP  r^Ragreldicn  Note: ENP stands for the effective number of parties. The values are based on parties share of parliamentary seats. The scores are based on the author's calculations.  Both trend lines indicate the gradual move toward a more fragmented party system in India. However, unlike Sartori's framework, both the Rae and the LaaksoTaagapera indices take the relative sizes of parties into consideration. Another such classificatory device building on the relative size of parties is offered by Ware. While Ware's notions of predominant and two-party systems are essentially the same as those of Sartori, he breaks down the multiparty types into four distinct groups: two-and-a-half party systems (defined by the absence of a majority party, and the two largest parties' winning at least 80% of the seats such that the third party is large enough to hold the balance of power, i.e. to give one of the two largest ones a majority); systems with one  109  large and several small parties (defined by the presence of a large party winning at least 45% but not more than 50% of the seats); systems with two large parties (defined by two parties' wmning at least 65% of the seats such that no other party holds more than 14%); and even party systems (defined by the largest party wining not more than 45% of the seats such that the two largest parties win less than 65% of the seats in total). According to Ware's classification, the Indian party system has moved from predominance to twopartism in 1977, back to predominance from 1980 to 1989, and to an even party system since 1989, as Figure 3.5 illustrates. FIGURE 3.5: T H E E V O L U T I O N OF T H E INDIAN P A R T Y S Y S T E M IN T E R M S OF WARE'S CLASSIFICATION Period 1952-1977  1977-1980 1980-1989  1989-present  System type Predominant party system (The Congress Party formed every government in this period) Two-party system (The Janata Party formed an alternative majority government) Predominant party system (The Congress(I) formed both governments in this period) Even party system (In 1989, the largest party was the Congress(I) with 37.7% of the seats. The two largest parties were the Congress(I) and the Janata Dal, together they controlled only 64.2% of the seats. In 1991, the largest party was again the Congress(I) winning 42.7% of the seats. The two largest parties were the Congress(I) and the BJP, together they controlled 64.8% of the seats. Finally, in 1996, the largest party, the BJP won 29.6%, and the two largest parties, the BJP and the Congress(I), only 55.4% of the seats.  In sum, in slight contrast with the conventional periodization proposed in the Indologist literature, comparative typologies of political party systems identify four rather  110  than six stages in the development of the Indian party system. The next section will seek to explain why the transition from one to another type of system took place.  3.2  EXPLAINING P A R T Y S Y S T E M C H A N G E : T H E C O N V E N T I O N A L WISDOM OF T H E INDOLOGIST L I T E R A T U R E The conventional explanation for the period of Congress dominance is provided  by Rajni Kothari  235  who claims that the Indian National Congress became a dominant  party because it was the exclusive legatee of the national consensus forged during the nation's struggle for Independence from Britain's colonial rule. As the Congress was the main protagonist of the Independence movement it managed to appeal to and forge political capital from the patriotic sentiment of the electorate. In addition, Kothari also emphasizes the importance of the Congress Party's organizational and distributive superiority vis-a-vis its competitors. According to him, the Congress put together winning coalitions of diverse communities and social groups both horizontally and vertically from the local to the national level through an elaborate network of faction chains. The Congress used its organizational machinery not only to mobilize at election times but more importantly to distribute particularistic benefits to groups whose electoral support it targeted. Chibber and Petrocik have noted that India's social complexity and territorialized cleavage structure encouraged the Congress to form a broad-based electoral coalition.  236  In particular, Chibber and Petrocik have found that in each state of the country a different combination of social groups formed the electoral base of the Congress Party. What held these very different social groups together in one party was the flow of patronage and particularistic benefits that control over the institutions of government and administration ^Rajni Kothari, "The Congress 'System' in India", Asian Survey (1964), pp. 1161-73. ^Pradeep Chibber and John R. Petrocik, "The Puzzle of Indian Politics: Social Cleavages and the Indian Party System", British Journal of Political Science 19 (1989), pp. 191-210.  Ill  afforded the Congress Party. In contrast, opposition parties had no such access to patronage which prevented them from aggregating into one party at the national level as they did not have the glue, i.e. patronage, that could have held their diverse social coalitions together. Thus, opposition parties remained locally confined, each representing its particular social group or social coalition in the state or region where it commanded influence. As a direct consequence of this electoral strategy appealing to a diversity of groups, the Congress had to become more of a pragmatic than an ideological party. This ideological and organizational pragmatism supposedly allowed the Congress to respond to changes and challenges in a flexible manner. At the same time, fearing absorption by the dominant party, the opposition parties were forced to became tighter-knit, more disciplined and ideological than the Congress. In turn, ideological rigidity further perpetuated the dominance of the Congress Party by making electoral coordination among the ideologically opposed non-Congress parties very difficult.  237  The late 1960s marked the beginning of what Manor calls the deinstitutionalization of the political and party systems in India. Electoral and party politics had supposedly become much more volatile than before as a result of the increased personalization and centralization of authority in the hands of Indira Gandhi. Numerous studies amassed evidence to show the organizational decline of the Congress. Elections 238  allegedly became plebiscitary affairs in which the personality of the leader, read Indira Gandhi, weighed more heavily on voters' behavior than substantive policy programs. It is this emphasis on the volatility of the electorate and the shifting alignments of social groups that became the conventional paradigm of studying Indian elections and the  ^James Manor, "Parties and the Party System", in Atul Kohli, ed., India's Democracy: An Analysis of Changing State-Society Relations (Princeton: Princeton University Press, 1988), pp. 62-5. ^ T o r example, see Stanley Kochanek, "Mrs. Gandhi's Pyramid: The New Congress", in Henry Hart, ed., Indira Gandhi's India (Boulder: Westview Press, 1976).  112 evolution of the party system. The drop in the Congress Party's share of the popular vote in the 1967 election was generally attributed to the demographic change affecting the electorate. By 1967 two decades had passed since Independence which meant that there were many new voters with little association with and memory of the pre-Independence times. For them, the Congress Party was just one of the many political options to choose from rather than a symbol of national pride and assertion. According to Morris-Jones this change in the demographic structure of the electorate created a market polity with increased competition among political parties.  239  The reassertion of power by the Congress Party in the 1971 election has been conventionally explained to be the result of a massive wave in favor of the leader of the Congress Party, Indira Gandhi, and her populist program to eradicate poverty (garibi hatao). A scholarly consensus emerged to consider this election as a critical realigning one as minorities and the underprivileged masses flocked to the Congress en masse responding to Indira Gandhi's populist appeal. The populist program was supposed to have created a nation-wide wave in favor of the Congress. It was argued that a new model of politics was emerging in which the distribution of patronage at the local level lost its relevance as a factor determining electoral behavior to wave-making issues.  240  The Congress Party's loss in the 1977 election has been said to be the result of an anti-Congress wave. Allegedly, the electorate turned against the party because of the excesses of the Emergency regime imposed by the Congress government in 1975. In  "^yndraeth Morris-Jones, "From Monopoly to Competition in India's Polity", in W. Morris-Jones, ed., Politics Mainly Indian (Madras: Orient Longman, 1978), pp. 144-59. Also see, Rajni Kothari, "Continuity and Change in the Indian Party System", Asian Survey (1970), pp. 937-48. See Davey Hampton, "Polarization and Consensus in Indian Party Politics", Asian Survey 12 (1972), pp. 701-16; Stanley Heginbotham, "The 1971 Revolution in Indian Voting Behavior", Asian Survey 11 (1971), pp. 1133-52; Oliver Mendelsohn, "The Collapse of the Indian National Congress", Pacific Affairs 51 (1978), pp. 41-65; Lloyd I. Rudolph, "Continuities and Change in Electoral Behavior: The 1971 Parliamentary Election in India", Asian Survey 11 (1971), pp. 1119-32. 240  113  particular, the Muslim and the Scheduled Caste communities were reported to have abandoned the Congress party in significant numbers.  241  The 1980 election resulting in a Congress revival has been explained to be the result of a pro-Congress wave after the Janata failed to provide a stable government. Just as the 1971 and the 1977 polls, this election has also become regarded as a realigning critical election because the Congress supposedly forged a new electoral coalition that allowed it to win.  242  The 1984 election has not been regarded a critical election because it  continued the dominance of the Congress Party. The dominant issue that was said to decide in favor of the Congress was the assassination of Prime Minister Indira Gandhi. Supposedly, voters flocked to the Congress out of sympathy for the son of the deceased leader, Rajiv Gandhi, who took over the leadership of the Congress. In the post-1989 period national issues have been argued not to be relevant any longer as no political party has managed to sweep the national polls. Instead, elections have been claimed to be decided by regional issues and shifts in favor of one or another political party. This is exactly the reason why the post-1989 party system has seen the rise to prominence of a significant number of small regionally concentrated parties.  243  The 1991 election is a partial aberration in this regard. Rudolph has argued that in the absence of a favorable electoral wave and a de-institutionalized party machinery the Congress(I) was heading towards an electoral defeat in the first pre-assassination phase of that election that, if continued unabated, would have been worse than the one the Party had suffered in 1977. However, as a result of the outpouring of the electorate's sympathy Myron Weiner, India at the Polls: The Parliamentary Election of 1977 (Washington, D C : American Enterprise Institute for Public policy research, 1978). Harold A. Gould, "The Second Coming: The 1980 Elections in India's Hindi Belt", Asian Survey 20 (1980), pp. 595-616; Myron Weiner, "Congress Restored: Continuities and Discontinuities in Indian Politics", Asian Survey 22 (1982), pp. 339-55. Rekha Saxena, "Party System in Transition", in M . P. Singh and Rekha Saxena, eds., India's Political Agenda: Perspectives on the Party System (Delhi: Kalinga Publications, 1996), pp. 72-77; M . P. Singh, "From Centrist Predominance to Polarized Pluralism: The Post-1989 Party System in India", Punjab Journal of Politics 16(1992). 241  242  243  114  with the terrible demise of its leader, the Congress(I) Party improved its performance in the remaining second and third phases significantly. The conventional interpretation of the evolution of the party system has acknowledged that the electoral performance of the Congress Party has never been evenly distributed across the regions of the country as Figures 3.6 and 3.7 illustrate.  FIGURE 3.6: T H E R E G I O N A L B R E A K D O W N OF T H E CONGRESS P A R T Y ' S E L E C T O R A L P E R F O R M A N C E , 1952-96  MM  1  ii  1952 19571962 1967 1971 19771980 1984 1989 19911996 Election year •  Hindi Heartland • I West  South  •  Northwest  East  Source: Authors' calculations based on data from Butler, Roy and Lahiri, India Decides, various tables Figure 3.6 shows the percentage of the seats that the Congress Party won in a given region as a share of all seats won by the party in the particular election year. The data highlight that historically the Congress Party's electoral stronghold was the Northern region of the country known as the Hindi Heartland.  244  Although the Congress was  For the purposes of examining the regional pattern of the Congress Party's performance five geo-political regions have been identified. The first, the Hindi Heartland, consists of the states of Bihar, Delhi, Hirnachal Pradesh, Madhya Pradesh, Rajasthan and Uttar Pradesh. The second, the South, includes the states of Andhra Pradesh, Karnataka, Kerala and Tamil Nadu as well as the Union Territories of Lakshadweep and Pondicherry. The third region is the Northwest containing the states of Haryana, Punjab and Jammu and Kashmir as well as the Union Territory of Chandigarh. The fourth region is the West 244  115  defeated badly in this region in the election of 1977 it quickly regained its strength there by the next poll. Since 1989, however, the Hindi Heartland has no longer constituted a reliable electoral terrain for the Congress Party. As the Figure shows, the Congress has become stronger in the South than in the Heartland over the last three electoral cycles since 1989. Prior to 1989, the Congress Party won 43.4% of its total seats in the Hindi Heartland on average, and only 27.3% in the South including the election of 1977. Excluding that election, the figures slightly change to 49% in the Hindi Heartland and 22.5% in the South. Since 1989, however, the percentage of its total seats that the Congress Party has won in the Hindi Heartland has dropped below this historical average to a mere 25.7% whereas the corresponding figure for the South has risen to 39.7! The region-wise performance of the Congress Party may also be examined by looking at the actual percentage of seats that the Party won in each region, as in Figure 3.7. Again, the data clearly reveal that of the two larger regions, i.e. the Hindi Heartland and the South respectively, the Congress has been fundamentally weakened in the former as of late. Whereas prior to 1989 the Congress managed to win an average of 75.8% of the seats in the Hindi Heartland, not counting the election of 1977, since 1989 the Party has been able to win only an average of 20.5% of the seats there. In the South the Party's post-1989 performance has not changed much from its historical average. Prior to 1989 the Congress has won 61.4% of the seats in the South on average while in the post-1989 period this figure has slightly dropped to 60:5%. These statistics suggest that the overall national decline of the Congress Party since 1989 has been largely due to its electoral slippage in the Hindi Heartland, the region that alone accounts for almost half of all  consisting of the states of Goa, Gujarat and Maharashtra and the Union Territories of Daman and Diu, and Dadra and Nagar Haveli. Finally, the fifth region is the East encompassing the states of Assam, Arunachal Pradesh, Manipur, Meghalaya, Mizoram, Nagaland, Sikkim, Tripura and West Bengal as well as the Union Territory of Andaman and Nicobar Islands.  116  parliamentary seats.  FIGURE 3.7: T H E CONGRESS PARTY'S E L E C T O R A L P E R F O R M A N C E IN T H E REGIONS, 1952-96  3.3  TESTING C O N V E N T I O N A L WISDOM Kothari's thesis explaining the dominance of the Congress Party in the early post-  Independence period with reference to the legacy of the national consensus forged during the Independence movement offers an idiosyncratic explanation that does not help in understanding the evolution of the party system through the various stages in a consistent manner. According to this explanation the so-called Congress system could only last as long as the political generation to whom the Independence movement was a salient electorally mobilizing issue was alive and constituted a majority of the electorate. As Figure 3.8 shows, the 1967 general election was the last one in which the majority of the eligible electorate would have personal memories of the pre-Independence era and, perhaps, association with the Congress Party.  117 FIGURE 3.8: T H E A G E S T R U C T U R E OF T H E INDIAN E L E C T O R A T E A N D T H E E L E C T O R A L P E R F O R M A N C E OF T H E CONGRESS P A R T Y  1957  1962  1957  1971  1977  1980  1984  Election yesr % d a s 1f>in'47  %INCvte  • %IN3vlecr4LBtecl  Source: William G. Vanderbok, "The Tiger Triumphant: The Mobilization and Alignment of the Indian Electorate", British Journal ofPolitical Science 20 (1990), pp. 239.  Thus, if the 'national consensus' explanation were correct and consistent, then the Congress should have been dominant from 1952 through 1967 and should have experienced a secular decline in its electoral fortunes thereafter. However, as Figure 3.8 also reveals, there has been no strong association between the two variables. Although until the 1967 election the Congress' vote share, shown by the dashed line, seemed to follow the decline in the number of those voters who had been 16 years of age or older at the time of Independence, this relationship was interrupted in 1971, when the Party's vote share suddenly increased and disappeared entirely after 1977. In spite of the gradual decline in the size of the mentioned segment of the electorate, the Congress Party managed to increase its share of the popular vote in the 1980s. The relationship becomes even weaker if the Congress Party's vote share is  118  indexed against electoral turnout. This measure is more robust because it shows the overall level df Congress support amongst the entire electorate not only those who actually turned out to exercise their franchise. After all, the poor relationship between the two variables, if one relied only on the percent of the votes polled by the Congress Party, may be the result of the inability and/or unwillingness of those voters who were 16+ in 1947 to turn out to vote on election day. However, as the dotted line in Figure 3.8 shows, the electoral base of the Congress, in terms of the percentage of eligible enfranchised voters who voted for the party in the various elections, has not fluctuated much over the years. Whereas the standard deviation of the Congress Party's share of the votes cast between 1952 and 1996 is 5.48, the standard deviation of the Party's share of the total eligible electorate for the same years is 3.34. It is also interesting that in the 1967 and 1971 elections the Congress Party actually received the support of a larger segment of the total electorate than in the first two post-Independence elections. The parallel thesis that explains Congress dominance in the early stage of the party system with reference to the broad social-electoral coalition the party forged suffers from both circular reasoning and ad hocery. First, regarding the issue of particularistic benefits, the distribution of which supposedly held the Congress coalition together, it is not clear whether it was control over the institutions of government and the consequent distribution of the spoils of office that enabled the Congress to maintain its widely based electoral coalition or the other way around. Either way, however, this explanation would suggest that the Congress could lose electoral support only if it failed to deliver the goods to its supporters. However, it is hard to see why the party would do so. On the other hand, had the party consistently kept on delivering benefits to its constituents then its hold on office should have been perpetual. The problem of ad hocery is just as troubling. Although the social structure of the country may well encourage the formation of a broadly based party it is not at all clear  119  why there would be only one such party formed. If Chibber and Petrocik are correct in concluding that the Congress is just as diverse in terms of the social composition of its electoral coalition as the opposition taken en bloc, then it is difficult to see why the Congress would have managed to aggregate these divergent streams and groups in one coalition whereas the opposition has failed to do so. Again, as mentioned, the explanation offered is that the Congress, as the governing party, can distribute patronage and thus hold diverse interests together. At this point, however, the reasoning once again becomes circular. It suggests that being in office allows the Congress to forge a broad electoral coalition which in turn perpetuates the party's hold on office. However, the explanation does not reveal how the Party got into office to begin with. The conventional thesis that the post-Congress system evolution of the party system has been driven by volatile electorates responding to plebiscitary campaigns, issues and personalities has been convincingly debunked by Vanderbok.  245  He finds that  Contrary to the dominant model of Indian politics, which posits the Congress as a master manipulator of bloc voting, and despite a large literature that portrays regular crises and realignments in the Indian electorate, circumstantial evidence does not indicate that any realignments have taken place, particularly if realignment is defined as an enduring shift in the propensity of identifiable groups to change their political orientation. Instead, there seem to be large, enduring and different groups that regularly mobilize in favor of the Congress or opposition banners in the various regions of the nation. The Congress is relatively consistent in its ability to mobilize and relatively unable to expand on its base. The opposition base, too, is stable but opposition parties experience comparatively wider swings in their ability to bring new voters into the system and then keep them over time. The result is a situation of contained volatility. 246  Vanderbok arrives at this conclusion by assessing the performance of the Congress Party vis-a-vis the opposition over time not in terms of the actual shares of the vote, and seats, they received, which has been the methodology followed in the conventional literature, but in terms of their share of the mobilized vote, i.e. the parties' Vanderbok, "The Tiger Triumphant"  'Ibid, p. 259.  120  share of the support of the total eligible electorate. Figure 3.9 clearly shows that the Congress Party's electoral performance over time has been far less volatile than that of the opposition. The standard deviation of the Congress Party's aggregate mobilized vote across the eight elections that Vanderbok considers is 2.8 in contrast with the 4.6 for the opposition. Figure 3.9 also sheds doubt on the existence of both the Indira wave in 1971 and the anti-Indira wave in 1977. In neither of these elections did the Congress Party's electoral performance, in terms of vote mobilization, deviate from the normal trajectory. Thus, the extent to which these elections can be called re-aligning is seriously questioned by Vanderbok's finding. Curiously, it is the 1984 election, which the conventional literature does not classify as a critical or re-aligning election because it merely continued the dominance of the Congress Party, that appears to deviate from the historical pattern.  FIGURE 3.9: T H E A C T U A L A N D MOBILIZED V O T E OF T H E CONGRESS A N D T H E OPPOSITION PARTIES, 1952-96  % O  70i 60 50 40 30 <  *  i .  20 > 10 1  n  1952  •  »— '  —  •  1957  1962  1967  1971  •——'  1977  1980  1984  Election year INC mobilized vote OPP mobilized vote  INC valid vote - * - OPP valid vote  Source: Vanderbok, "The Tiger Triumphant," pp. 242-3  Looking at the performance of the opposition overtime, the picture changes significantly. As the data in Figure 3.9 indicate the opposition's electoral performance  121  has been significantly more volatile than that of the Congress. On the basis of these findings, Vanderbok plausibly concludes that "variations in Congress party performance have been due less to what the Congress does than to the ability of the opposition to mobilize voters. It is opposition volatility of mobilization that is the driving force behind shifts in the party's electoral fortunes."  247  Vanderbok also refutes the conventional wisdom that the Congress' electoral stronghold has been in the Northern Hindi Heartland. Figures 3.10 and 3.11 compare the Congress Party's and the opposition's performance in the Hindi and the non-Hindi speaking regions over time. FIGURE 3.10: T H E A C T U A L A N D MOBILIZED V O T E OF T H E CONGRESS A N D OPPOSITION PARTIES IN T H E HINDI H E A R T L A N D , 1952-84  192  1957  192  1957  1971  1977  BecrJcnyea"  -•- ll\Cnrr±ilized\Gte INCvelidvate GFPnxb'lized\jCte -*- CFPvelidvcte Source: Vanderbok, 'The Tiger Triumphant," p. 247.  Vanderbok, "The Tiger Triumphant..." p. 246.  1930  19B4  122 FIGURE3.il: T H E A C T U A L A N D MOBILIZED V O T E OF T H E CONGRESS A N D OPPOSITION PARTIES IN T H E NON-HINDI REGIONS, 1952-84  10 1952  1957  1962!  1967  1971  1977  1980  1984  Election year INC mobilized vote OPP mobilized vote  INC valid vote - * - OPP valid vote  Source: Vanderbok, "The Tiger Triumphant," p. 247.  The charts clearly demonstrate that the Congress has actually been stronger electorally in the non-Hindi speaking areas. Across the eight elections from 1952 through 1984, the Congress Party's mean mobilized vote share is 20.75% in the Hindi states as opposed to 25.62% in the non-Hindi states as shown in Table 3.4. The Party's share of the valid vote in the non-Hindi areas also exceeds that in the Hindi-speaking region by an average of 1.75%: whereas the Congress has mustered 42.5% of the valid votes cast in the Hindispeaking regions it has won 44.25% of the valid vote in the non-Hindi states. Finally, the electoral support that the Congress has mobilized in both regions over time has shown much less volatility than that mobilized by the opposition. The standard deviation of the Congress Party's mobilized vote in the Hindi speaking region is 4 as opposed to the opposition's 6.4. In the non-Hindi areas the opposition's mobilized support has been more stable as suggested by a standard deviation of 3.5. However, this is still higher than the standard deviation of the Congress' mobilized vote in the same region, which is 3. Although Vanderbok's contribution is indispensable in that he has highlighted  123  how misperceived and misinterpreted the development of the Indian party system has been, it does not offer an alternative explanation that can be used to provide an consistent account of the history of the party system. 3.4  C O M P A R A T I V E THEORIES OF P A R T Y S Y S T E M C H A N G E  One of the most influential, and controversial, theories on party systems has been offered by Maurice Duverger. According to Duverger, the number of parties in a party system 248  is determined by the electoral formula, i.e. the rules that convert the votes electors cast into legislative seats. He claimed in what later William Riker  249  called Duverger's Law  that "the simple-majority single ballot system favors the two-party system".  250  In what  Riker called Duverger's Hypothesis, Duverger states that "the simple-majority system with second ballot and proportional representation favors multipartism."  251  Duverger offers two reasons why the simple plurality system may favor a twoparty system. The first is what he calls the mechanical factor, which simply means the automatic elimination of all but the winning candidate in an electoral district. Since there is only one winner, which is indicated in the more apt term "single-member simple plurality system", in a given district, all the votes that have been cast for candidates other than the winner are wasted. The second reason is the psychological factor which means that voters seek to avoid wasting their votes and will support candidates that stand a reasonable chance to win. Since there is only one seat to be won, these factors favor the consolidation of the electorate behind two candidates. This kind of voter behavior is called tactical or strategic voting.  252  In exposing the effect of proportional representation and the majority run-off  Maurice Duverger, Political Parties (New York: Wiley, 1954) William H . Riker, "The Two-Party System and Duverger's Law: An Essay on the History of Political Science," American Political Science Review 76 (1982), 753-66. Duverger, Political Parties, p. 217. 248  249  250  Ibid., p. 239. ^Gary W. Cox, Making Votes Count: Strategic Coordination in the World's Electoral 251  Systems (Cambridge: Cambridge University Press, 1997), p. 13.  124  systems on the format of the party system, Duverger follows a similar logic. Proportional representation by definition reduces the amount of wasted votes by rewarding every candidate in proportion to the number of votes he/she mustered. In majority run-off systems there is a single winner, similarly to simple plurality and in contrast with proportional representation, that is determined on a second ballot after weak candidates are eliminated from the race on the basis of their returns on the first ballot.  253  This,  formula encourages the fragmentation of the party system because candidates are eliminated only after the results of the first round are known. Duverger's' theory has been tested, critiqued and refined time and time again.  254  An important extension of Duverger's theory has been related to the problem of crossdistrict linkage. Duverger recognizes that his theory about strategic voting in simple plurality systems is limited to local, i.e. district-level dynamics. In other words, the logic of strategic voting that he posits predicts the consolidation of bipartism only at the district but not the national level. According to him, national bipartism automatically follows the evolution of district-level bipartism because "increased centralization of organization within the parties and consequent tendency to see political problems from the wider, national standpoint tend of themselves to project on to the entire country the localized two-party system brought about by the ballot procedure."  255  Leys' criticism of Duverger led him to offer an alternative argument about the ^There may be variations on the actual rules of entry into the second round. For example, in French elections, only the top two candidates proceed from the first round and the winner, by definition, will have a majority on second ballot. In contrast, in Hungary, candidates winning at least 15% of the votes in the first round can enter the second round where the winner needs to secure only a simple plurality of the votes. ^ F o r an excellent overview of the various streams of criticism that Duverger's Law and Hypothesis have been subjected to, see Gary W. Cox, Making Votes Count, chapter 2. The classic empirical study on the relationship between electoral formulae and party system format is Douglas Rae, The Political Consequences of Electoral Laws (New Haven: Yale University Press, 1971). Also see, Andre Blais and R. K. Carry, "The Psychological Impact of Electoral Laws: Measuring Duverger's Elusive Factor", British Journal of Political Science 21 (1991), pp. 79-93. ^Duverger, Political Parties, p. 228.  125  nature of strategic voting and linkage.  256  According to him, strategic voting does occur, as  Duverger predicts, but it is in favor of the two parties that have the greatest chance to win nationally rather than in favor of the two locally most viable parties, or candidates.  257  Another argument offered by Sartori tries to explain cross-district linkage by the degree to which the party system is structured and is composed of highly developed mass parties.  258  Although Sartori himself remains vague on these terms, it is plausible that the  national two-party system will tend towards bipartism if mass parties field candidates in every district in each of which the logic of strategic behavior drives the system towards local bipartism. Cox offers an instrumental explanation of linkage. According to him, individual politicians have an incentive to link and run on the same party label across districts because of the advantage that the economies of scale provide them in the attainment of some important task in the legislature. Cox offers four potential versions of this argument. First, politicians may link when motivated by the prospect of controlling 259  government policy.  260  The party history literature that distinguishes between internally  and externally created parties supports this point. The birth of internally created parties was the result of legislators' desire to affect government policy in parliament. In order to enhance their future prospects and continuity such parties had an incentive to organize and mobilize the electorate on a wide-scale. Similarly, externally created parties, such as the ones born out of nationalist, independence or trade union movements etc., were the ^Colin Leys, "Models, Theories and the Theory of Political Parties," Political Studies 7  (1959), 127-46. ^Ibid, p. 142. ^Giovanni Sartori, "Political Development and Political Engineering," in John D. Montgomery and Albert O. Hirschman, eds., Public Policy (Cambridge: Cambridge University Press, 1968), p. 281, 293. "''Cox discusses the incentives that the office of a strong president and that of a strong prime minister provide for linkage separately. Therefore, he mentions five potential versions of his linkage argument, see Making Votes Count, p. 186. However, for the purposes of highlighting the logic of the argument there is no need to separate the two. ^Cox, Making Votes Count, p. 187.  126  extension of an already existing network of linkage. That such movements were converted into political organizations is plausible because of their interest in capturing government office and being in direct control of policy rather than just pressuring it from the extra-parliamentary arena. Second, linkage may be driven by parties' pursuit not of government policy in general but the most important executive office in particular, for example the Presidency or the Prime Ministership.  261  The more powerful these offices are, the stronger the  procedure by which these offices are filled, the more strongly legislative and executive offices are linked the greater the incentive for politicians seeking to capture these offices will be. Third, politicians may want to link up and create large national parties when there is an upper tier of seat allocation that rewards broadly based parties. Finally, the 262  dynamics of campaign finance may also encourage linkage as donors will want to provide funds to parties that stand a realistic chance to affect national policy in their favor.  263  Similarly, campaign finance regulations may explicitly encourage linkage by  tying the provision of public funds to electoral largesse. As for voters, Cox claims that they have to weigh both the local and the national competitiveness of their preferred party. No matter how competitive nationally a party is, voters will not have an incentive to vote for it in a given district unless its candidates stand a reasonable chance to win there. "The reason is that a vote that cannot change a local outcome cannot a fortiori change the national outcome, and thus has no instrumental value either at the local or the national level."  264  Furthermore, it may make  sense for voters to support a party that nationally is not but locally is competitive.  'Cox, Making Votes Count, pp. 187-92.  Ibid., p. 192. 'Ibid, p. 192. 'Ibid, p. 183. l  127 The reason is that electing a nationally hopeless party's candidate increases the probability of a hung parliament (and thus perhaps the probability of participation in a coalition government), and decreases the probability that one of the nationally competitive parties will secure a majority, both things that some instrumentally motivated voters may want to do. 265  The strongest criticism of Duverger's theory has been put forth by the sociological school of political scientists. The central argument of these critics is that institutional forces play a minor role in shaping the party system compared to the importance of social cleavages. The essential thrust of the social cleavage theory is that 266  cross-cutting cleavages increase while overlapping cleavages reduce the number of political parties. If the cleavages that run across a society overlap with one another, meaning that the same social groups would consistently be on the same side of each fault line, then the cleavage structure is simple and the drive towards multipartism is restrained. On the other hand, if the politically salient social cleavages cut across one another it will be impossible to maintain politically homogeneous large social groups. For example, in countries with sharply politicized and cross-cutting religious and linguistic cleavages religious communities will be divided by the different language their members speak while the linguistic communities will be divided along denominational lines. According to the cleavage theory of party systems, such social structure provides a fertile ground for multipartism as there are multiple well delineated hunting grounds that parties can target to build an electoral base on.  267  ^'Cox, Making Votes Count, p. 183. ^The sociological literature on party systems is voluminous. For some of the foundational works, see Jean Blondel, Comparing Political Systems (New York: Praeger, 1972); Klaus von Beyme, Political Parlies in Western Democracies (New York: St. Martin's Press, 1985); J. G. Grumm, "Theories of Electoral Systems," Midwest Journal ofPolitical Science 2 (1958), 357-76. Leslie Lipson, "Party Systems in the United Kingdom and the Older Commonwealth: Causes, Resemblances, and Variations," Political Studies 7 (1959), 12-31; John Meisel, Cleavages, Parties and Values in Canada (Beverly Hills: Sage Publications, 1974); Seymour Martin Lipset and Stein Rokkan, eds., Party Systems and Voter Alignments: Cross-National Perspectives (New York: Free Press, 1967) M7  F o r a long time India has been considered to be an exception to the social cleavage  128  3.5  A T E S T OF T W O THEORIES If Duverger's Law held the key to understanding the transition from a pattern of  majority to hung parliaments in India, then there should have been a change in the electoral formula that could precipitate such a transition. However, while there were changes in India's electoral formula, they took place a long time before the transition to a fragmented party system and pattern of hung parliaments. The single-member simple plurality electoral system was not really in place until only the third parliamentary elections in 1962 because of the presence of a number of double-member districts in the first two polls. In 1952, of the 489 national parliamentary seats 172 were filled by the top two candidates in the 86 double member constituencies and 3 by the top three finishers in the single triple member constituency in the state of West Bengal. In 1957, the number of double-member constituencies was increased to 91 and the number of parliamentary seats filled by candidates from such districts jumped to 182. At the same time, the triple-member seat in Bengal was abolished. In December 1960, the Indian Parliament passed the Abolition of Two-Member Constituencies Act which divided each of the existing 91 double-member districts into single-member ones. From then onwards every parliamentary seat has been awarded according to the singlemember simple plurality formula. Therefore, since there has been no change in the electoral formula towards a more proportional allocation of seats among political parties, Duverger's law cannot account for the fragmentation of the Indian party system. In addition, it is also worth mentioning that according to Duverger's law, the fragmentation of the party system should have been the greatest in the first two national elections  theory in light of the fact that the country's enormous degree of social heterogeneity was not reflected in a corresponding degree of party system fragmentation. However, in an important study Chibber and Petrocik have shown that this puzzle of Indian politics could be explained by the heterogeneous nature of the electoral coalitions that have supported the dominant Congress Party. See, Pradeep Chibber and John R. Petrocik, "The Puzzle of Indian Politics: Social Cleavages and the Indian Party System", British Journal of Political Science 19(1989), 191-210.  129  because the number of non-single member districts was the largest then. However, as Figure 3.4 illustrates, the party system was actually the least fragmented in those two foundational elections. According to the social cleavage theory, the fragmentation of the Indian party system must be the result of fundamental changes in the structure of politically salient cleavages in Indian society which should be reflected in changes in electoral behavior. The conventional measurement of such change is the Pedersen index of volatiliy which calculates the difference in parties' share of the popular vote between two elections summed across the entire party system.  268  Thus, if the fragmentation of the Indian party  system were caused by rising rates of electoral volatility, which reflects the changing cleavage structure of the country, then the Pedersen index should show rising values towards the end of the 1980 and the beginning of the 1990s. Figure 3.12 clearly shows that this has not been the case. Whereas electoral volatility was on the rise between 1967 and 1977, it has been declining steadily ever since. Therefore, the fragmentation of the party system as of 1989 cannot be attributed to the kind of change in electoral behavior that the social cleavage theory suggests.  268  The Pedersen index is the following:  where p. is the percent of the vote received by party i in election t, and n is the number of parties competing in the two elections considered. See, Mogens N . Pedersen, "On Measuring Party System Change: A Methodological Critique and a Suggestion", Comparative Political Studies 12 (1980), pp. 387-43; and "The Dynamics of European Party Systems: Changing Patterns of Electoral Volatility", European Journal of Political Research 7 (1979), pp. 1-26. According to Bartolini and Mair's revision, if two or more parties merge between elections then the vote total of the individual parties in the premerger election is compared with the vote total of the new party in the post-merger election. Similarly, i f a party splits into fragments between two elections then the united party's share of the vote is compared with the vote total of the individual fragments in the post-split election. See, Stefano Bartolini and Peter Mair, Identity, Competition and t  Electoral Availability:  The Stabilization of European Electorates 1885-1985 (Cambridge:  Cambridge University Press, 1991).  130 FIGURE 3.12: E L E C T O R A L V O L A T I L I T Y IN INDIA, 1952-96  10 9 8  £  7  1  6  UJ  3  ra  E  5  21  n  1  •  .  .  1952-571957-62 1962-671967-71 1971-77 1977-80 1980-841984-89 1989-91 1991-96  Election years  3.6  STRATEGIC V O T I N G A N D L I N K A G E In his seminal article published in 1976 William Riker sought to formalize the  theoretical conditions under which the single-member simple-plurality electoral formula produced a two-party system, as predicted by Duverger. Riker found that Duverger's law could work only either if party competition were not structured by ideology, or the center party had no chance of winning; or the center party had a reasonable chance to win but it was an expected first loser in the race, and it only faced one other party in the contest. On the other hand, if either the center party were an expected first loser but the party system contained more than just two parties; or the center party were the expected winner of the race then instead of consolidating around two parties, the system will be perpetually fragmented at the district level. There are three fundamental assumptions behind Riker's analysis with regard to the behavior of individual voters. First, similarly to Duverger, Riker expects that voters will not want to vote for parties that have no reasonable chance to win. Second, however,  131  Riker allows for the presence of sincere voters, i.e. voters who are committed to voting for the party of their first choice no matter what. Finally, Riker also allows for the presence of disillusioned voters who will desert the largest party because of their disappointment when the party they supported does not provide them with the benefits and rewards they had anticipated. Thus, assuming a hypothetical party system with three parties, A , B , and C, where C may stand for any further number of parties above 3, the preferences of voters, N , may be schematized as follows. N ,: A>B>C  N ,: B>A>C  Al  N  N  BI  • A>C>B A2  N  : OA>B  Cl  'B>OA  N  B2  : OB>A.  C2  In the absence of an ideological dimension Riker finds that a two-party system is guaranteed to emerge even when all three, or n, parties enter the race in the first place. For example, if the first election produces an outcome A>B>C then in the second election voters N  will cast their ballots for A while voters N  c 2  will vote for B given that C has  no chance to win. As Cox notes, this may not happen immediately when B and C muster very similar shares of the popular vote because voters NCI,2 may believe that the party of their first choice may stand a chance to win. However, the relative difference in the two party's strengths should become clear over time eventually resulting in the emergence of a two-party system.  269  At the same time, A may suffer from disillusioned  voting and may lose voters either to B or C. Again, once it becomes clear whether B or C is larger the extent of disillusioned voting will change: only those N voters will desert A  the party of their first choice whose second preference party has a realistic chance to win. Thus, if C is very small without any hope of winning, then disillusioned voting will originate solely from among voters N  . In sum, permanent multipartism may not emerge  under the single-member simple-plurality formula when ideology is absent from the party competition.  'Cox, Making Votes Count, p. 32.  132  The situation is quite different when ideology becomes an important factor to consider. Assuming that there is a single ideological dimension along which both parties and voters locate themselves such that A is in the center, B is on the left and C is on the right, both disillusioned and strategic voting are made more difficult because voters will now have to consider not only the relative sizes of but also the relative ideological distance between the parties. In a nutshell, the presence of an ideological dimension implies that voters will not vote for the party of their second choice, whether out of disillusionment or strategy, unless it is located next to the party of their first choice along the ideological spectrum.  270  If the ranking of parties after an election at time T is O B > A that is the center party is the smallest, then the effect of disillusioned and strategic voting is not different at all from what they are like in a non-ideological context. Disillusioned voters N C j can flee to A while N C will have to stay with C because the ideological distance they would have 2  to travel to vote for the party of their second preference is too far. At the same time, the effect of disillusioned voting is reduced, or even canceled out, because strategic voting by both N A , and NA„ is unlimited. The outcome is the eventual desertion of the small J  1  2  center party resulting in a two-party system. When the center party, A, is the middle party then the party system cannot transform into a two-party system unless it started out as such. When multiple parties enter the race at time T and the result is B>A>C then both disillusioned and strategic voting are half prohibited. Disillusioned votes can only be cast by NBj whereas strategic votes can only be cast by NCy Since both would benefit the center party, it is quite likely that A would grow and might even win the next election at T However, both B and C would continue competing because a part of their electoral base remained intact. In contrast, if the system started out as a two-party system such, for example, that B>A at 270  Riker, ""The Number of Political Parties," p. 101.  133  T , then bipartism would remain intact. Strategic voting would not take place at all, while disillusioned voters N B would flee to A. 1  Finally, if A emerged as the largest party after the race at T , further supposing that B>C, then the result would be perpetual multipartism. Since the largest party is in the center there is no Umitation on disillusioned voting, however, strategic voting is only half allowed . Thus, the smallest party, C, would not disappear and multipartism would continue. Furthermore, even if the system started out as a two-party race between A and B such that A>B at  a new party C would certainly enter the race at T because there 2  would be a body of disillusioned voters seeking to flee A towards the right. According to Riker, this last scenario was a perfect representation of how the India party system operated during the Congress-dominant era. The largest party, the Congress was in the center which meant that disillusioned voters could always leave it both to the left and to the right. However, the disillusioned vote that the center party loses gets divided between the left and right parties, B and C above, making it highly probable that the center party will still remain the largest of all. In addition, the center party further benefits from the strategic vote that some supporters of the hopeless party would cast for it. In sum, Riker's model accounts for the perpetual fragmentation in the India party system dominated by the large center party as the endogenous outcome of a model based on the principles of strategic and disillusioned voting. This essentially district-level model has significant implications for the phenomenon of majority and hung parliaments in India. According to the last theoretical scenario described, the center candidate or party of every district has a great probability to win. Therefore, it makes perfect sense for them to link across the districts and form one party in order to ensure that it would emerge the largest political force in the parliament after all races are completed in the districts. At the same time, a similar aggregation of the left and right parties may take place with much less likelihood simply because it would not yield any significant benefits. Even if all left and all right parties aggregated and  134  linked across the districts, so long as the center candidates are the largest at the district level, they can at most emerge as large opposition parties at the national level. This, however, is hardly an incentive to incur the high transaction costs of linkage in an executive dominated parliamentary system that gives the opposition little effective say in policy making. Therefore, the model can quite plausibly account for the asymmetrical aggregation that has characterized India's party system for the most part: a consolidated center facing a fragmented left and right. Riker's model suggests that any change resulting in the decline of the center party at the district, and consequently at the national, level has to come from exogenous sources. Since the key to the defeat of the center party at the district level is to ensure that disillusioned voting, which is fleeing the center, is maximized, while strategic voting, which is benefiting the center, is kept to a minimum at the same time, the off-center parties have to engage in a difficult coordination exercise in each district to attain these twin ends. In other words, to defeat the center party in the districts, the off-center parties have to ensure that they consolidate behind the candidate or party that has the greatest likelihood to attract the disillusioned vote from the center while having the smallest likelihood to suffer from strategic desertion. This requires both strategic entry in and abstention from the race, in other words, the formation of an ends-against-the-center coalition in as many districts as possible.  271  The incorporation of strategic entry, a variable entirely ignored by Riker, in models of rational voting under plurality rule was first proposed by Downs and has been enhanced and refined by a number of researchers. See, Anthony Downs, An Economic Theory of Democracy (New York: Harper and Row, 1957); Joseph Greenberg and Kenneth Shepsle, "The Effect of Electoral Awards in a Multiparty Competition with Entry", American Political Science Review 81 (1987), 525-37; Thomas Palfrey, "Spatial Equilibrium with Entry", Review of Economic Studies 51 (1984), 139-156; Timothy Feddersen, Itai Sened and Stephen G. Wright, "Rational Voting and Candidate Entry under Plurality Rule", American Journal of Political Science 34 (1990), 1005-16; Martin J. Osborne, "Candidate Positioning and Entry in a Political Competition", Games and Economic Behavior 5 (1993), 133-51; Gary W. Cox,Making the Votes Count... pp.151 78. The term "ends-against-the-center coalition" is borrowed from Gary W. Cox, Making 271  the Votes Count, p. 239.  135  To illustrate this point in terms of the schematic model, suppose that B and C arrived at an agreement to coordinate the nomination of their candidates in the various districts with the defeat of the center, A, candidates in as many district as possible in mind. Further, suppose that the distribution of voters, according to their preference ranking of parties, at T l were as follows NAji  20%  NBj!  10%  NC^  15%  NA,:  20%  NB •  25%  NC •  10%  2  2  2  and that the disillusioned voter group were NA1. Given these pieces of information, it makes sense for C to abstain from the race. By so doing, the subjective ranking of the candidates by the six groups of voters will change at T2 as shown below. N ,: A>B  N  N  N  Al  • A>B A2  While N  C 1  Bl  : B>A  N : A>B ci  : B>A  N  B2  : B>A  C2  will cast a strategic vote in favor of A as he/she would have done anyway, N  Q  is now forced to make a decision between either supporting B, which is ideologically costly, or abstain, which, however, will only help A who is N ' s least preferred option. C2  If NC2 decides to vote strategically, rather than abstain, then B will end up having 45% of the vote, while, however, A will also benefit from the strategic vote cast by N C I in its favor which may increase its vote total to 55%. At the same time, since the NA1 group of voters is disillusioned they are likely to switch their support to B, which will now have a total of 65% of the vote and will win the contest. Had the disillusioned camp of voters been NA2, disillusioned voting probably would not have taken place at all because these voters would not have had a candidate, or party, to cast their vote for given the abstention ofC. In sum, in order to build an effective ends-against-the center coalition the offcenter parties have to have sophisticated intelligence about the distribution of voter support. More specifically, they have to know the expected size of the disillusioned and  136  strategic vote relative to one another. The threshold vote share above which A can not be defeated is 50%. Evidently, should candidate A be expected to receive more than half of all the votes cast in a district he or she may no longer be beaten rendering the cooperation of off-center parties futile. However, so long as A is expected to receive less than 50% of the vote there is always a theoretically possible level of coordination at which the offcenter parties may defeat it. The degree of coordination that the off-center parties need to reach in order to at least tie with the center party in a given district is given by the following formula:  272  Ax^l-Ax ) 1  where A i s the percent of the vote, expressed in decimal form, received by candidate A l  x  in district X in an election held at t point in time. For example, if A is expected to receive 35% of the vote in election T l in district X , then the non-A parties have to engage in a 54% degree cooperation, meaning that the largest non-A party in the district should receive no less than 54% of the non-A vote, which, of course will translate into 35.1% of the total vote. As Figure 3.13 shows, connecting all the points at which the largest non-A candidate can tie with A yields the Tie Curve. At any point above the curve A is defeated and at any point below the curve A wins. The slope of the curve suggests that with every percentage point increase in A's share of the vote electoral coordination among the nonA candidates will become more and more difficult. Similarly, the weaker candidate A is expected to be in a given election in a given district, the less difficult electoral coordination among the non-A candidates will be because the largest non-A candidate will not have to have as much of the total non-A vote concentrated in him/her.  This formula is based on the Index of Opposition Unity constructed by Butler, Lahiri, and Roy. See, David Butler, Ashok Lahiri and Prannoy Roy, A Compendium of Indian Elections (New Delhi: Arnold Heinemann, 1984). 272  137  FIGURE 3.13: THE TIE CURVE  Non-A unity  Whereas the formation of effective ends-against-the-center electoral coalitions determines the outcome of center-dominated multicandidate races at the district level, the aggregate outcome at the national level depends on the cross-district linkage that these off-center parties manage to attain. If the off-center parties ensure that their victorious candidates run on the same party label in a majority of the districts then a majority parliament will be elected with the largest off-center party controlling the majority of the seats, i.e. the center-dominated party system is transformed into a two-party system. However, if the off-center parties fail to run their successful candidates on the same label in a majority of the districts then a hung parliament will emerge and the party system is transformed into a competitive multiparty format. In essence, one can think of three possible electoral outcomes in a centerdominated multiparty system depending on the effectiveness of the off-center parties'  138  cooperation at both the district and the cross-district levels, as illustrated in Figure 3.14.  FIGURE 3.14: POSSIBLE TRANSFORMATIONS OF A C E N T E R - D O M I N A T E D M U L T I P A R T Y SYSTEM Effective E A C C at district Ineffective E A C C at  Effective E A C C at cross-  district level  Ineffective E A C C at  cross-district level  level  district level  Majority parliament dominated by largest offcenter party. Two-party system. Hung parliament. Competitive, even multiparty system.  Majority parliament dominated by center party. Predominant party system. Majority parliament dominated by center party. Predominant party system.  Note: EACC stands for ends-against-the-center coalition As the Figure suggests, a truly effective ends-against-the-center coalition requires the cooperation of the off-center parties not only at the district but also at the supradistrict levels. If the latter fails, then depending on the off-center parties' success to form an effective anti-center coalition at the district level, either a hung parliament or a majority center-dominated parliament may emerge. Finally, even though the off-center parties may link effectively across the districts, they will be unable to dominate the parliament because of their failure to build an effective electoral coalition at the district level. While ineffective cross-district linkage by the off-center parties reduces the likelihood of the election of a majority parliament, it increases the national viability of each off-center party individually, provided of course that they have engaged in effective coordination at the district level resulting in the displacement of a sufficiently large number of centrist candidates. This is very important because as recent advances in both formal theoretical and empirical research have shown, voters take into consideration not only the local viability of candidates but also the national viability of their parties when deciding whether and for whom to cast a strategic vote. Cox notes two ways in which national considerations may affect voters'  139  assessment of a candidate's local viability. On the one hand, the national viability of a candidate's party may serve a tie-breaking function when the voter is not quite decided about his/her preference between two candidates. On the other hand,  another way in which lack of national viability may contribute to a party's fall is via a drying up of contributors. Here, the logic may well be more national than local: favor-seeking contributors have no reason to contribute to locally hopeless candidates but neither do they have much reason to contribute to locally viable candidates from nationally unimportant parties. In order to get a return on their investment, the candidate must not only win a seat but also be able to do something with that seat once in parliament. 273  To incorporate this insight in the Rikerian model, suppose that the country had five districts: P,Q, X , Y and Z. Further suppose that B and C, located on the left and the right of the ideological dimension respectively, had formed an electoral coalition against A, that was in the center, in election T in each of the five districts such that C fielded the coalition candidate in P and Q while B ran the coalition candidate in X , Y and Z. In each district the coalition was successful in defeating A with the exception of X where A prevailed. Following the election, B and C formed a coalition government. In the subsequent election T the ends-against-the-center coalition of B and C is no longer together and both parties run their own candidates separately in each of the five districts. Although C had been a third party candidate in each district prior to election T , it would no longer be considered such because of its strong performance nationally in election T and because voters do care about the outcome of the elections in the country as a whole as well as the local circumstances. In districts X , Y and Z, where B had been the coalition candidate in election T  2  and where C had abstained, both N in election T . N  C J  and N  C 2  voters will have an incentive to vote for C  voters will once again have the option to vote sincerely for C, whereas  Cox, Making the Votes Count... pp. 195-6. Also see, Richard Johnston, Andre Blais, Henry E . Brady and Jean Crete, Letting the People Decide: Dynamics of a Canadian Election (Stanford: Stanford University Press, 1992), 225. 273  140  N  voters who otherwise had voted for A can also vote sincerely for C because the  proven national viability of C no longer makes them feel that their sincere vote for C is wasted. In districts P and Q where C had been the coalition candidate and where B had abstained, the same dynamics would take place with the result that both N voters will support B in election T whereas N 2  m  and N  K  and N C 2 voters will vote for C.  The overall outcome of these dynamics will be the evening out of voter support among the three candidates. It is important that the only candidate who loses votes is A and he/she loses precisely that strategic vote that is now cast sincerely for the former third party, C, whose national viability has now made it a rational choice for voters to support. Finally, the weaker electorally A becomes the less coordinated the non-A forces have to be in each district in order to defeat it. Thus, as A loses strategic voters whose support it had enjoyed in the past, on top of whatever disillusioned vote it also loses, it becomes easier for the ex-coalition partners to defeat it even though they are no longer explicitly united and coordinated. For these processes to take place, it is very important that C, the originally unviable third party candidate, proved relatively successful in election T . For example, suppose that the coalition were successful in defeating A only in districts X , Y and Z. If so, then in election T with B and C running separately in each district again, only B will 2  remain regarded as a viable competitor against A. While the re-entry of C in these districts in election T2 will provide an incentive for N their sincere vote for them, N  C 2  voters to return to C and cast  voters will not perceive C any more viable nationally  than it used to be before! Thus N  will keep supporting A and the Rikerian equilibrium  is restored. Similarly, the re-entry into the race by B in districts P and Q in election T would 2  allow N  and N  B 2  voters to support their favorite candidate once again. Although C may  have come close to defeating A in these districts and there may have been a good chance that he/she would actually succeed in doing so at the next polls, the re-entry of B and the  141  transfer of N election  B 2  voters from C to B would certainly prevent this. Had C defeated A in  in these districts and had C formed a coalition government with B, the party's  national viability would have improved which in turn would have balanced the movement of  by retaining the support of both N  failure, C will lose N  Q 2  to B and N  C 1  C 1  and N  C 2  voters. However, because of its  to A in both P and Q with the likely result that the  Rikerian equilibrium is restored. The foregoing theoretical discussion suggests that the direction of the transition from a center-dominated party system is the result of the effectiveness of the coordination among the off-center parties at the district and the cross-district, or national, levels. Whereas the emergence of a two-party system reflects effective off-center coordination at both levels, a movement towards a more even multiparty system reflects the fact that offcenter parties failed to coordinate their strategies effectively across the districts, however, at the district level they coordinated well enough to deprive the center partyfromwinning the election. 3.7  A STRATEGIC COOPERATION E X P L A N A T I O N OF C H A N G E S IN T H E INDIAN P A R T Y S Y S T E M It has been well established in the literature on the Indian party system that the  Congress Party has been in the center of the competitive political space. While this assertion has been developed primarily on the basis of analyses of the party's policy resolutions, campaign strategies and alliances with other parties, Huber and Ingelhart have recently provided an empirically more objective support to it. According to the findings of a survey of country experts, the Congress Party stands squarely in the middle of the left-right spectrum, as Table 3.1 shows. In addition, as mentioned already, Riker himself deemed that the Indian party system was a natural analogue for his model of a center-dominated multiparty system. If so, then the explanation of the emergence of hung parliaments in post-1989 India must lie, as per Figure 3.14, in the asymmetric coordination of the non-Congress  142  parties at the district and the supra-district levels: whereas they are able to defeat the Congress locally they are unable, or unwilling, to form a majority alternative to it in the national parliament. This stands in sharp contrast with the outcome of the 1977 election when the non-Congress, i.e. off-center, parties succeeded in coordinating both locally and nationally, resulting in a majority parliament dominated for the first time by an off-center party, the Janata. T A B L E 3.1: T H E IDEOLOGICAL L O C A T I O N OF INDIA POLITICAL PARTIES  PARTY  LEFT-RIGHT SCORE 2.22 2.64 4.5 AIADMK 5.67 TNC(I) 5.8 DMK 6.5 BJP 8.18 CPI(M) CPI JD  IUML  9.00  Note: The locational figures refer to the parties' position on a 10-point scale. The higher the score the closer the party is to the right. Source: Huber and Ingelhart (1995: 98-99).  In four of the eleven general elections, specifically in 1977, 1989, 1991 and 1996, the non-Congress parties succeeded in arriving at a high enough degree of coordination that was sufficient to defeat the Congress candidates in a majority of the districts as shown in Table 3.2. The Table compares the degrees of opposition coordination that would have been necessary just to tie with the Congress in the majority of the districts in each election with the actual degree of coordination that the opposition parties arrived at. Clearly, in each of these four elections the non-Congress parties managed to coordinate their electoral strategies far above the degree of coordination that would have been necessary to just tie with the Congress. Accordingly, they were rewarded for their cooperation as the Congress Party lost its historical majorities in these elections.  143  T A B L E 3.2: E L E C T O R A L COORDINATION AGAINST T H E CONGRESS P A R T Y , 1962-1996 Election year 1962 1967 1971 1977 1980 1984 1989 1991 1996  R D C - % Congress seat % Congress vote R D C A D C A D C 44.7 80.8 67 13.8 73.5 40.8 68.9 67 1.9 54.4 . 43.7 77.6 71 6.6 67.9 34.5 52.7 90 -37.3 28.4 42.7 74.5 65 9.5 65.1 48.1 92.7 74 18.7 76.6 39.5 36.5 28.8  65.3 57.5 40.4  77 66 62  -11.7 -8.5 -21.6  36.3 42.7 25.8  Notes: Thefirstcolumn shows the percentage of the popular vote obtained by the Congress Party in each election; RDC stands for the Required Degree of Coordination that is necessary for opposition parties to reach in order to at least tie with the Congress Party at a percentages of the vote the Party won in each election; ADC stands for the Actual Degree of Coordination that opposition parties arrived at in each election; the third column shows the difference between the Required and the Actual Degrees of Coordination; the last column shows the percent of the seats the Congress Party won in each election. Source: Butler, Lahiri and Roy, India Decides, various tables and author's calculations. FIGURE 3.15: T H E P E R F O R M A N C E OF T H E CONGRESS P A R T Y A N D T H E E F F E C T I V E N E S S OF NON-CONGRESS COORDINATION  144  Figure 3.15 further illustrates the positive relationship between the Congress Party's electoral performance, in terms of the percentage of the seats won and the closeness to which the opposition parties approximated the level of coordination at which the dominant party could be defeated. As suggested earlier, the weaker the centrist Congress Party, in terms of its share of the votes cast, is expected to be, the smaller the degree of electoral coordination that the non-Congress parties have to arrive at in order to defeat the Congress candidates. Thus, although the actual degree of electoral coordination by the non-Congress parties increased by only 3 percent between 1984 and 1989, this small increase was effective enough given the large decline in the percentage of the vote that the Congress received between these elections. On the other hand, in 1977 the opposition parties reached such a high level of electoral coordination, at 90%, that it would have allowed them to tie with the Congress even if the dominant party had increased its share of the vote up to 47.4! The sudden decline in the percentage of the vote mustered by the Congress Party in both the 1977 and 1989 elections made the task of the opposition significantly easier. Whereas in 1971, the last election prior to 1977, the Congress won 43.7% of the popular vote it only managed to win 34.5% six years later. Similarly, in 1984, the last election prior to 1989, the Congress won 48.1% of the votes cast in contrast with a mere 39.5% five years later in 1989. The dynamics of how effective ends-against-the-center coalitions work can be examined by looking at a couple of actual examples. For instance, as Table 3.3 shows, Pilibhit constituency displayed a classic Rikerian scenario in 1971 as the centrist Congress prevailed over a divided off-center opposition.  274  The fact that the Congress'  vote share was well below 50% suggested that the party's candidate was potentially defeatable. In the 1977 election, the opposition parties, that had run candidates separately 2 7 4  Each of the following constituency tables are based data in Butler, Lahiri and Roy,  India Decides.  145  against the Congress six years before, were partners in the Janata coalition and ran a united candidate on the B L D ticket. As expected, this resulted in a spectacular victory of the coalition over the Congress candidate. That the Congress' vote share dropped and the B L D candidate's vote share rose far above the combined vote share of the INCO, BJS and B K D in 1971 means that there was a large scale disillusioned desertion of the Congress candidate. T A B L E 3.3: PUmHTT C O N S T I T U E N C Y IN 1971 A N D 1977  Party  1971  1977  Congress Congress(O) BJS BKD BLD  39% 24.70% 19.10% 7.90%  19.70%  Others  9.30%  71.30% 9%  Similarly, Mahasamund constituency also displayed a classic Rikerian scenario in the 1984 election, however, five years later, the BJP and the Janata Dal had an agreement not to run candidates against one another in the district. According to the inter-party arrangement the Janata Dal ran a candidate against the CongressfJ) in this constituency and, although barely, managed to win.  T A B L E 3.4: M A H A S A M U N D CONSTITUENCY IN 1984 A N D 1989  Party  1984  CongressfJ) 57% BJP 23%  1989 43.30%  Janata Party 10.80% Janata Dal 45.70% Others 9.20% 11%  146  The recovery of the Congress Party in 1980 was inherent in the way its defeat was brought about three years before, i.e. in the merger of the opposition parties. Because the five parties had merged and run candidates on the same ticket, there was only one party, the Janata, that derived gains from the defeat of the dominant party in terms of enhanced national viability and visibility. However, once the party disintegrated, it became unclear to the voters which Janata fragment, if either, had or could have the potential to be an alternative to the Congress in the next election. In terms of the convention used earlier, the distribution of voters' preference among the parties involved went through the following changes as a result of the split : 275  N ^ : JP>INC  —> N ^ : JP>JPS>INC —_> N  : JP>INOJPS JP2  —-> N ^ : JPS>JP>INC ~ - > N ^ : JPS>INC>JP  N  J N C  : I N O J P —> N  m c i  : INC>JP>JPS  -->  N  W C 2  : INOJPS>JP  The split of the Janata Party had three important effects on the dynamics of voting. First, N  and N  voters had an incentive to vote for their sincere first choice  candidates. With the Congress(I) being in the center and the coalition no longer forcing them to make an ideologically costly choice, these voters could now switch their support to their most preferred candidates. Second, the split also created an incentive for N  and  voters to vote strategically in favor of their second choice, the Congress, because  they could not be sure whether their first choice candidate would be strong enough to defeat their least liked candidate, i.e. either that of the Janata Party (Secular) or that of the In the following JP stands for Janata Party, JPS for Janata Party (Secular), and INC for Congress.  275  147  Janata Party. Third, the split left the N  and N ki "I \* X  voters unaffected. t i N \ « » i  On balance, the overall effect of the split of the Janata Party and the subsequent re-entry of itsfragmentswas a net gain of votes by the Congress and an inefficient division in the formerly united Janata vote. Thus, the Rikerian Congress-dominant multiparty system was restored in 1980. A good illustration of this is provided by the constituencies of Agra and Bombay North. In the 1977 election in Agra the united Janata candidate, contesting on the symbol of the B L D , inflicted a devastating defeat on the Congress candidate by reducing the latter's vote share by 32.6% compared with the previous election. However, three years later, in 1980, when the Janata party was no longer together and both its successor fragments, the Janata Party and the Janata Party (Secular) ran candidates in the district on their own the Congress(I) candidate triumphed by increasing his share of the vote by 8.4% while the combined vote share of the two Janata candidates fell by 10.6%. T A B L E 3.5: A G R A CONSTITUENCY Party  1977  1980  1989  1991  26.50% 34.90% 33.90% 16.60% 70.30% Janata Party 30% Janata Party (Secular) 29.60% Janata Dal 54.70% 30.60% BJP 46.20% Others 3.20% 5.50% 11.40% 6.60% Congress BLD  In contrast, in the 1980 election in Bombay North constituency, of the two Janata fragments only the Janata Party ran a candidate against the Congress(I). As expected, the absence of a Janata party (Secular) candidate did not allow the Congress (I) to take advantage of the split that the Janata Party had suffered. Although the Congress(I) candidate made some gains in terms of votes and the Janata Party candidate lost compared to the B L D candidate's vote in 1977 the Janata candidate still prevailed.  148 T A B L E 3.6: B O M B A Y NORTH CONSTITUENCY  Party  1977  1980  Congress BLD  31.60% 67%  37%  Janata Party Others  1.40%  51.40% 11.60%  The Congress Party has failed to repeat its comeback ever since its debacle in the 1989 general election, even though the ends-against-the center coalition that had defeated it broke up in 1990. According to the theoretical model, just as the party's recovery in 1980 was inherent in the manner in which it had been defeated in 1977, so has the party's post-1989 decline been inherent in the manner in which it was defeated in the 1989 election, i.e. by an opposition coalition that was not only effective in terms of defeating the dominant party but also propelled its members to national system-wide prominence and viability. In terms of the abstract convention, the distribution of electoral preferences following the break-up of the 1989 coalition and the subsequent re-entry of the race by its members separately can be described as below : 276  NJW"  1  J D O  3  * ) ^  -"-> >  N  m c  : INC>JD(BJP)  N  J M ( B J P I )  : JD(BJP)>BJP(JD)>lNC  NjwnW JD(BJP)>rNC>BJP(JD)  > N ^ : INC>JD>BJP  •> N  r  . INC>BJP>JD  INC2  For the sake of simplicity, only the coalition formed by the Janata Dal and the BJP is described here. Although this is somewhat of a simplification of the complex reality of an Indian election, it does convey the main point. In addition, the coalition of these two parties did provide the backbone of the broader anti-Congress alliance in 1989. Depending on whether the Janata Dal or the BJP ran the coalition candidate in the district in 1989 N ^ or N is the applicable voter to consider. 276  277  JTJ  HJP  149  The break-up of the 1989 coalition had two important impacts on the dynamics of voting. First, once the coalition had broken up and its ex-members entered the electoral fray on their own in 1991, N  voters, for whom supporting the coalition candidate  in 1989 was already costly in ideological terms, had an incentive to vote sincerely for the candidate fielded by the party of their first choice. These were the voters whose first choice party had to abstain from running a candidate in the particular district according to the coalition agreement and whose least preferred alternative was the Congress Party and its candidate. Second, however, this re-balancing of electoral support among the non-Congress parties and candidates was also accompanied by the retention of the support of N  J D 2 ( B J p 2 )  voters by the Janata Dal and the BJP respectively. Because the dissolution of the coalition did not entail a change in the identity of these parties voters had very clear information about their relative standing and position in the party system after the election. The drastically improved overall national standing of the ex-coalition partners individually provided an incentive for those voters who would have voted for them sincerely but had voted instead for Congress for strategic reasons earlier to reconsider their behavior and vote sincerely for their first choice. In sum, since N  J D 2 ( B J P 2 )  voters no longer perceived  their sincere vote wasted they had an incentive to support the candidate of their first choice. On balance, the distribution of electoral support among the three sets of parties in the average district was the following under the altered incentives:  I N C N ^ N  JD:  N J D I  N INC2  ^JD2  BJP:N  B J P  N 1N  BJP2  This scheme clearly shows that the distribution of voters among the three parties' candidates in the districts has become much more balanced than before and that the  150  Congress no longer enjoys a structural advantage over the others. With this sort of distribution of electoral support short-term fluctuations and swings may be especially potent determinants of election outcomes. Good illustrations of the dynamics described above are provided by the electoral results in the districts of Agra, Rajapur and Begusarai respectively. In 1989 in Agra, as shown in Table 3.5, the Janata Dal ran the coalition candidate in the district. The Dal candidate was successful while the Congress(I) candidate lost 9.2% of his share of the vote compared with the last election held in 1984. Two years later, however, in 1991, when both the Janata Dal and the BJP entered the race in the district the Congress(I) was unable to recover its loss and slid further down in terms of the share of the popular vote it received. In Rajapur constituency, in 1989, the anti-Congress(I) electoral coalition also brought about the victory of a Janata Dal candidate. However, two years later, unlike in Agra, it was not the BJP but its small, nationally hardly viable, ally the Shiv Sena that ran a candidate against both the Janata Dal and the Congress(I). As expected, the Congress(I) did rather well and won the competition because it was not squeezed between the candidates of nationally viable parties. T A B L E 3.7: RAJAPUR CONSTITUENCY Party  1977  Congress BLD  32.10% 27.60% 43.60% 43.30% 67.90%  Janata Party Janata Dal Shiv Sena Others  1980  1989  1991  71.70% 52.90% 24% 0.70%  3.50%  31.40% 1.30%  Finally, the electoral results in Begusarai constituency of Bihar state illustrate the impact of the split in the Janata Dal. In 1989 the Dal fielded the victorious coalition candidate against the Congress(I). Two years later, however, there were three separate  151  anti-Congress(I) competitors entering the race: the Janata Dal, the BJP, and the Samajwadi Janata Party candidates. As expected, this resulted in the victory of the Congress(I) candidate: had the Janata Dai's share of the electorate not been split between the Dal and the Janata Party in 1991, the Congress(I) could have been squeezed between the BJP and Dal. However, the split in the Dal and the subsequent participation of both of its fragments in the electoral race resulted not only in an inefficient division of the Dai's 1989 voter base but also in the movement of some voters, who had supported the Dal in 1989 and whose least preferred candidate was the BJP, to the Congress(I). Again, because of the spit, Dal voters could not be sure whether either of the Dal fragments would be in a strong enough position to defeat the BJP. Therefore, those of them whose second choice was the Congress(I) candidate had a strong incentive to cast a strategic vote for him/her. T A B L E 3.8: BEGUSARAI CONSTITUENCY Party Congress BLD Janata Janata Party Janata Dal Janata Party (S) BJP Others  1977 36% 44.50%  1980 60.10% 12.80% 22%  1989 42.90%  1991 51.30%  55.30%  41.10% 2.70% 4.10%  19.50%  5.10%  1.80%  0.80%  The same dynamics can be further confirmed by looking at the aggregate data in Table 3.9 which clearly reveals that the gains that the Congress(I) made in 1980 in terms of its vote share were concentrated in those districts of the country where both the Janata Party and the Janata Party (Secular) ran candidates. Curiously, the Party even lost votes in those districts where either just one or neither of the Janata fragments nominated a candidate. The 1991 and 1996 figures also confirm the explanation offered above. The Congress(I)'s decline in term of its votes share was much more pronounced in those  152  districts in both of these elections where both the Janata Dal and the BJP ran candidates. T A B L E 3.9: T H E R E C O V E R Y A N D DECLIN1E OF T H E C O N G R E S S ^ P A R T Y  Year  Congress vote change  Congress vote change in post-coalition districts  Congress vote change in the rest of the districts  1977-1980 1989-1991 1991-1996  8.20% -3.00% -7.70%  8.90% -2.00% -7.10%  -0.70% -1.00% -0.60%  Note: Post-coalition districts in row 1977-80 are those constituencies where both the Janata party and the Janata Party (Secular) ran candidates in 1980. In 1991 and 1996 post-coalition districts are those where both the Janata Dal and the BJP ran candidates. The last piece of the puzzle that needs to be accounted for is why the off-center parties have no longer found it in their interest to aggregate and form a majority alternative to the CongressfJ) Party at the national level. The survey of the theoretical literature earlier on suggested that the degree to which parties aggregate in a party system is determined by the institutional incentives that either encourage or discourage them to do so. According to Cox, for parties to aggregate they have to be motivated either to capture a strong executive office, or to capture control over government policy, or to benefit from the economies of scale in upper tier seats allocations, or perhaps to benefit from campaign finance regulations. While there is no evidence suggesting that any of these potential sources of party aggregation may have changed, the insertion of the Tenth Schedule in the Constitution in 1985 has clearly raised the costs of party aggregation, especially so for off-center parties. As described, the Tenth Schedule makes party aggregation difficult by imposing heavy penalties on splits. Unless one third of a party's legislative contingent agrees to a split, representatives seeking to change their party affiliations have to stand for reelection in their constituencies. Knowing this, political parties have an incentive to merge only if they have a reasonable expectation that their unity in the legislature would last, however, the maintenance of such unity is especially difficult when political groups of diverging ideologies are attempting to merge.  153  To put it simply, the Tenth Schedule of the Constitution acts as a deterrent against party aggregation, especially against the aggregation of the off-center parties, making it more rational for parties to maintain their separate identities and engage in coalition bargaining in the legislature with one another. Furthermore, it has also been shown, that failure to coordinate across the districts has actually helped the non-Congress(I) parties to force the Congress(I) in a perpetual non-dominant status since 1989. Thus, the emergence of hung parliaments and an even multiparty system in India can be accounted for in terms of the rational choices that the off-center parties have made. In particular, there have been two choices involved. First, the off-center parties have learned that forming a single party may give them only a one-time legislative majority, provided they hit the right degree of coordination at the district level, whereas they can deprive the center party of its electoral dominance on a more permanent basis if they maintain their own separate identities. Second, the Tenth Schedule of the Constitution also provided an institutional disincentive against the merger of parties with different ideological predispositions. 3.8  CONCLUSION This chapter has sought to explain the transformation of the predominant party  system in India from a comparative perspective. In so doing it has shown that neither the mainstream Indologist nor the mainstream comparative explanations of party system change can offer an adequate and consistent account for the dependent variable. However, this does not have to lead to resorting to an explanation based on spatial and temporal idiosyncrasies, as it has been the case in much of the literature on Indian elections and party system. The chapter has demonstrated that William Riker's model of rational voting can be extended by findings of the more recent rational choice theoretical literature to provide a solid account for the decline of the Indian National Congress (I) Party and the subsequent transformation of the party system of India.  154  C H A P T E R FOUR: G O V E R N M E N T F O R M A T I O N This chapter will seek to provide a consistent account for government formation in India by evaluating the predictions and hypotheses offered in the comparative literature. First, it will be shown that coalition theories predicting the formation of minimum winning coalition governments are entirely inapplicable to the Indian cases. Second, the chapter will show that none of the models seeking to specify the conditions under which cabinets of various types are most likely to form can explain the formation of the Indian cabinets consistently either. Finally, the chapter will demonstrate that by incorporating constraints, both exogenous and endogenous, in the analysis most of the discrepancies between theoretical predictions and the actual empirical observations will disappear. The final part of the chapter will accordingly explain government formation in India as a series of rational choices made by individual political actors under macro- and microscopic constraints.  4.1  MINIMUM WINNING COALITION THEORIES  The classic theories of political coalitions put forth by Gamson  278  and Riker  279  predict the formation of coalitions of a minimum winning size. These theories are developed in the context of the theory of simple games which is explained in some detail in Appendix 2. Both Gamson and Riker assume that political actors are office-seeking, that they have perfect information about one another, that the coalition formation game is played by n persons, and that it is zero-sum with no side-payments allowed. In such games Riker predicts that participants will create coalitions just as large as they believe  William A. Gamson, "A Theory of Coalition Formation", American Sociological Review 26 (1961), p. 376. 278  William Riker, Theory of Political Coalitions (New Haven: Yale University Press 1962).  279  155  will ensure winning and no larger. This has become called the minimum winning 280  coalition theory. The reason why rational office-seeking actors would not want to accommodate superfluous members in the coalition, i.e. members without whose inclusion the coalition would still be winning the game, lies in the assumption that the game is zero-sum. Because the spoils of office, or the prize of the game, is fixed, the more coalition members there are the smaller the share of the prize that each member gets allotted. Thus, political actors will not want to share the prize with more actors than what is barely necessary for the coalition to win the game. In any coalitional situation, i. E . in games without a single winner, there may be multiple minimum winning coalition combinations of actors. For example, in a game of four actors A , B, C and D with weights {3,3,2,2} and a decision-making rule of majority, which means that the threshold of winning is 6, there are three minimum winning coalitions: A B , A C D , and BCD. The first coalition has a total weight of 6, the second and the third have a weight of 7 each. No member can be removed from any of the three coalitions without depriving the coalition of its winning status. According to both Gamson and Riker, in such situations it will always be the coalition with the smallest weight that will be formed, in this case AB. The minimum winning theory of coalitions was developed further by Leiserson, Axelrod and De Swaan who incorporated the idea of policy-seeking into the set of assumptions. The most influential of these has been Axelrod's thesis which posited that cooperative behavior is the function of the conflict of interest among actors. The less conflict of interest there is, the more likely is cooperative behavior.  281  With specific  reference to political coalitions, Axelrod predicted that the less conflict of interest there would be in a coalition the more likely that it was going to be formed. For Axelrod, conflict of interest in political coalitions is measured by the degree ^''Riker, Theory of Political Coalitions, p. 32. ^•Robert Axelrod, Conflict of Interest (Chicago: Markham, 1970), p. 167.  156  of dispersion of members of the coalition in a uni-dimensional ordinal policy-space. Since the smallest dispersion will be found in coalitions formed by actors that are connected along the dimension, Axelrod further predicts that coalitions whose members are connected along the single dimension will be formed. Finally, because members of such coalitions are also office-seekers and they want to control the branches of government they would want to form a wining combination. In sum, Axelrod proposed 282  that office- and policy-seeking rational political actors will seek to build minimum connected winning coalitions. T A B L E 4.1: CHARACTERISTICS OF INDIAN G O V E R N M E N T S , 1952-97 Lok Sabha  Prime Minister  1952-1957 1957-62 1962-1967  J. Nehru J. Nehru J. Nehru (until 1964) L.B. Shastri (until 1966) I. Gandhi I. Gandhi  1967-71  1971-1977 1977-1979  I. Gandhi M . Desai C. Singh  1980-1984 1984-1989 1989-1991  I. Gandhi R. Gandhi V.P. Singh C. Shekhar  1991-1996 1996-1998  N . Rao A.B. Vajpayee H.D.D. Gowda I.K. Gujral  Prime Minister's party Congress Congress  Number of parties in government 1 1  All Congress  1  Congress, after 1969 Congress  1  54.4%, after 1969 42.7%  1 2 2  67.9% 56.1% 28.0%  1 1 4 1  65.1% 76.6% 26.8% 10.7%  1 2 13 13  45.7% 32.6% 25.1% 25.1%  (R) Congress ( R ) Janata Party Janata Party (Secular) Congress ( I ) Congress ( I ) Janata Dal Samajwadi Janata Congress ( I ) BJP Janata Dal Janata Dal  'Axelrod, Conflict of Interest, p. 170.  Seat % of all parties in government 74.4% 75.1% 73.1%  157  Since both these theories predict the formation of coalitions of a minimum winning size they are inconsistent with the Indian cases. Of the sixteen government that were formed in India between 1952 and 1997, six were multiparty coalition governments, however, none of these were of a minimum winning size, as Table 4.1 indicates. Technically speaking, the only minimum size winning coalitions were the ones formed by the Congress and the Congress(I) whenever the party won a majority of the seats in the Lok Sabha. All other governments were either undersized minority governments, coalition or single-party, or an oversized coalition as in the case of the coalition of the Janata Party and the Akali Dal in 1977. Had the Akali Dal chosen not to participate in the coalition the Janata Party would still have been a winner.  4.2  P A R T Y S Y S T E M CHARACTERISTICS A N D G O V E R N M E N T TYPES The single most comprehensive theory that seeks to account for the type of  cabinet formed on the basis of various characteristics of the party system is offered by Dodd. Dodd argues that the type of government that is formed is a function of the degree of information certainty that parties can obtain about each other and their willingness to bargain with one another. Dodd claims that these two variables are determined by three 283  characteristics of the legislative party system: its fragmentation, its polarization and its stability. As Figure 4.1 shows, Dodd theorizes that undersized, read minority, governments are expected to form in legislatures that are polarized, fragmented and unstable because in such assemblies parties will find it difficult to obtain reliable information about one another's preferences and strategic, which in turn reduces their willingness to bargain with each other. With regard to polarization, it is important to note that conventional wisdom about Indian political parties has it that parties are not ideologically disciplined and ^Lawrence C. Dodd, Coalitions in Parliamentary University Press, 1976).  Government (Princeton: Princeton  158  therefore explanations of party behavior based on policy-seeking and ideological motivation are irrelevant in the Indian context.  284  The unfortunate consequence of this  conventionally accepted thesis has been the absence of research about the ideological location of political parties in India. As scholars have failed to find evidence that parties professing a leftist orrightistideological predisposition would either consistently coordinate their strategies with parties of a similar ideological outlook or pursue ideologically based policies once in office, the consensus that ideology is absent in Indian party politics has become unquestionably accepted.  FIGURE 4.1: DODD'S HYPOTHESES A B O U T CABINET F O R M A T I O N Party system traits Fragmented Polarized Unstable  Information certainty uncertainty  Willingness to bargain low  Cabinet to be formed undersized  Fragmented Polarized Stable  certainty  high  almost minimum winning  Defragmented Polarized Stable  certainty  high  almost minimum winning  Defragmented Depolarized Stable  certainty  high  minimum winning  Fragmented Depolarized Unstable  uncertainty  high  oversized  However, as mentioned in the previous chapter, Huber and Inglehart have recently found that Indian parties do care about ideology and they can be located on a left-right spectrum in a meaningful way. As Table 4.2 shows, using Huber and Inglehart's data allows for a fairly sophisticated calculation of the extent of party system polarization. However, since the data about the main parties ideological location are only  ^See, Paul R. Brass, "Coalition Politics in Northern India," American Political Science Review, (1968), pp. 1174-91, and Hampton Davey, "Polarization and Consensus in Indian Party Politics," Asian Survey, 12 (1972), pp. 701-10.  159  available for the 1990s they cannot be used to calculate changes in the degree of party system polarization over time. In any case, the Table does show that ideological polarization in the post-1989 legislatures has been rather moderate. T A B L E 4.2: IDEOLOGICAL POLARIZATION IN THE L O K S A B H A , 1989-1996 YEAR 1989 1991 1996  ELECTORAL  POLARIZATION LEGISLATIVE 2.5 2.6 2.5  POLARIZATION 3 2.6 3.7  Notes: To calculate the polarization scores, the following formula was used: I f (x-X) , where f is the vote/seat share of the 7-th party, expressed in decimals, at point t in time; x is the left-right score of the i-th party; X is the mean left-right score of the party system. For comparison of these scores with Western European data, see , Lane and Ersson (1994:185). 2  it  As for the entire five-decade history of the party system, including the pre-1989 legislatures, polarization scores can be obtained by re-operationalizing the variable. According to this alternative operationalization, polarization can be defined as the difference between the legislative weight of the center party or parties and the combined total weight of the parties of the left and the right, counting only those parties that control at least 3% of the legislative seats, as per Ware's rule-of-thumb.  285  The greater the  positive difference the less polarized the party system because of the strength of the center. However, the smaller and more negative the difference, the more polarized and centrifugal the party system due to the strength of the extreme parties. This way of operationalizing the variable is better suited for the Indian context because while no longitudinal data are available that would allow for the definition of the exact ideological location of parties in integral terms, impressionistic evidence does make it relatively easy to identify if there is a center party and which parties are to its left and right respectively. In formulaic terms, polarization is defined as  Alan Ware, Political Parties and Party Systems (Oxford: Oxford University Press 1996), pp. 158-9.  285  160  POL: pC(p>3%)- \pL(p>3%)+pK(p>3%)]  where p is the percentage of the seats held by a party; C is any party classified as being in the center; L is a party classified as being on the left; and R is a party classified as being on the right. Table 4.2 does suggest an overall positive relationship between polarization and the formation of minority governments in the Lok Sabha. As the bolded entries show, minority governments were formed when the polarization scores were at their lowest values. The sole outlier in this regard is the Janata party government formed in 1977. As the Table suggests, the Lok Sabha was the most polarized between 1977 and 1979, yet the Janata Party formed an oversized majority coalition with the Akali Dal.  286  T A B L E 4.3: POLA1RIZATION IN TTIE INDIAN PARIfY S Y S T E M , 1952-1996 Year 1952-1957 1957-1962  Weight of the center Weight of the Left 74.4 3.3 75.1 9.3  Weight of the Right Polarization NA 71.1 NA 65.8  1962-1967  73.1  5.9  3.6  63.6  1967-1969  57.2  12.5  20  24.7  1969-1971  40.4  12.5  40  -12.1  1971-1977  67.9  9.3  11.8  46.8  1977-1979  28.4  4  57.9  -33.5  1979-1980  28.4  12.7  38  -22.3  1980-1984  65.1  16.6  5.7  42.8  1984-1989  76.6  9.6  NA  67  1989-1990 1990-1991  363 363  32.2 32.2  16 16  1991-1996  42.7  173  22.1  -11.9 -11.9 33  1996-1998  25.8  24.5  29.7  -28.4  Even though there appears to be a positive association between the extent of polarization and the formation of minority governments in the Lok Sabha, it should be  ^The point that the Janata Party thus formed an oversized coalition as opposed to a single-party majority government is often ignored in most analyses of Indian political history.  161  emphasized that so far polarization has been used in a relative sense. Thus, to say that the Lok Sabha was the most polarized in 1977 is not the same as to say that it was highly polarized at the time. Since the value of the polarization index rangesfrom+100 to -100, one can speak of high and low levels of polarization when the value of the index is near the two extremes. Whereas the first Lok Sabha was indeed very close to the depolarized end of the spectrum, with a score of+71, there was no Lok Sabha anywhere near lower end of the range indicating high polarization. This interpretation of the finding confirms that since high polarization has not been a characteristic feature of the Lok Sabha, it cannot be the explanation for minority government formation as Dodd suggests. Figure 4.2 evaluates Dodd' hypothesis by considering the two variables of fragmentation and polarization together. FIGURE 4.2: FRAGMENTATION AND POLARIZATION IN THE LOK SABHA  162  Note: E N P P stands for the effective number of parliamentary parties. Once again, it is clear that at no point in time was the Lok Sabha highly fragmented and polarized, as the virtually empty lower right hand corner in the Figure shows. Curiously, at the time of the formation of the CongressfJ) minority government in 1991, the Lok Sabha was less polarized than at the time of the formation of the majority coalition by the Janata Party and the Akali Dal in 1977. The formation of this latter government provides a further rebuttal of Dodd's theory. According to him, oversized governments should form in depolarized, fragmented and unstable legislatures. However, the Sixth Lok Sabha in which this oversized majority coalition was formed was not highly fragmented at all. With regards to party system stability, Figure 4.3 shows that overall a high degree of volatility has characterized the Indian parliamentary party system. FIGURE 4.3: SEAT V O L A T I L I T Y IN INDIA, 1952-96 50 45 40-  .  / / / / /  35£  30 25  £  20 15  A  *\ \  /\ / \ / \ \ / \ \  10 5 nI 1 1 1 1 . 1952-57 1957-62 1962-67 1967-71 1971-771977-80 1980-84 1984-89 1989-91 1991-96  Election years  Source: Author's compilation from Butler, Lahiri and Roy, India Decides..., p. 67 and Election Commission of India, Statistical Report on General Elections, 1996 to the Eleventh Lok Sabha Volume 1 Table LS-05. 287  However, it is interesting that the volatility rate was rather small until 1969, when the  "'The calculation of the index is based on Pedersen's formula of volatility. The computation ignored the seat shares of parties classified as Others and Independents. See, Mogens Pedersen, "On Measuring Party System Change,"pp. 387-43.  163  first minority government was formed, and has become very small again, after two decades o f high levels and wide fluctuations, by the 1990s, a period in which only minority governments have been formed. This casts further doubt on the applicability o f Dodd's model to the Indian cases.  4.3  MACRO-POLITICAL REGIME CHARACTERISTICS The theory o f government types proposed by Luebbert offers a typological-  rational approach to understanding government formation.  288  Luebbert argues that  [a] typological approach would be one that distinguishes types o f political systems, the features o f which, in conjunction with a given constellation o f parties and preferences, will lead to certain types of outcomes. The key to such an approach is to find the chain of causality that runs from the distinctive features o f the political system type to the relative importance party leaders assign to their conflicting goals and from this system-prioritized set o f goals to a certain evaluation o f the prevailing set of policy preferences and then, in consequence, the government formation outcome. 289  In pointing out the need for a comparative historical understanding o f the choices that individual political actors make Luebbert is not calling for descriptive ad hoc studies. Quite the contrary, he wants the political system types and their characteristics to be specified prior to the analysis which should be "conducted within the context o f a deductive, rational choice model, that will enhance the theory".  290  In his typological model, Luebbert distinguishes political systems depending on (a) the degree o f their legitimacy, and (b) the participation o f opposition parties in consensus building, see Figure 4.4. A political system is considered to lack legitimacy, according to Luebbert, when there is an anti-democratic party in its party system o f such  ^ G r e g o r y M . Luebbert, "Coalition Theory and Government Formation in Multiparty Democracies", Comparative Politics 15 (1983), pp. 235-49; " A Theory of Government Formation", Comparative Political Studies 17 (1984), pp. 229-64; Comparative Democracy: Policymaking and Governing Coalitions in Europe andIsrael.(New York: Columbia University Press, 1986). l u e b b e r t , "Coalition Theory and Government Formation", p. 246.  "°Ibid, p. 247.  164  magnitude that its presence induces centrifugal competition.  291  As for the opposition, it  may either cooperate with the government, and then be a consensus builder, or may not cooperate, in which case there will be no consensus. The four system types that this typology yields are indicated in the Figure 4.4. FIGURE 4.4: L U E B B E R T ' S T Y P O L O G Y OF POLITICAL S Y S T E M S  Opposition is consensus builder High regime legitimacy L o w regime legitimacy  Opposition is not consensus builder  consensual party system competitive party system unconsolidated party conflictual party system system  The second element of Luebbert's model is the specification of the relationships that may exist between the most important party of the cabinet formation process, called the formateur's party, and the other parties. These relationships can be of three kinds: the parties' policy preferences may either be tangential, convergent and divergent. Tangential preferences mean that parties emphasize different issues that are not related to one another and therefore party leaders do not consider them incompatible. When parties emphasize the same issues such that their principles tend towards each others' then their preferences are convergent. Finally, when parties' preferences are mutually exclusive, their preferences and relationship are divergent, according to Luebbert. The third element of Luebbert's theory is the specification of the formateur's party's limit of tolerance in each political regime type in terms of these preference relations. In consensual democracies, the formatuer will include those parties in the cabinet whose preferences are tangential to or convergent with his. In conflictual democracies, the limit of tolerance is set by divergent preferences which means that all but the anti-systemic anti-democratic parties can participate in the formateur's cabinet. Finally, competitive democracies are divided into two subtypes: dominated and ^Luebbert, " A Theory of Government Formation", p. 230.  165  undominated. In dominated competitive democracies party leaders work on the assumption that a particular party cannot be left out of the cabinet. Obviously, the limit of tolerance in dominated competitive systems will be lower than in undominated competitive ones. The actual hypotheses concerning the formation of governments of different types are the following: •  Oversized governments will be formed in dominated competitive systems  •  Minimal winning coalitions, i.e. coalitions without superfluous members, will be formed in undominated competitive democracies  •  Minority governments will be formed in both consensual and conflictual democracies. In conflictual regimes minority governments are formed because parties are unable to cooperate to form majority coalition. In contrast, in consensual regimes there is no need for parties build a majority coalition. Luebbert's theory is not supported by the Indian cases. According to Luebbert,  minority governments should be virtually absent in competitive regimes. However, a brief look at the nature of the Indian political regime will convincingly show that India belongs to exactly this regime type. That the legitimacy of the regime has been high is indicated both by the historically high levels of voter turnout in elections, see Figure 4.5, and the absence of threat that the extreme parties would have posed to the political regime. The fact that both the extreme Right, the BJS later renamed to be the BJP, and the extreme Left, the CPI(M), have participated in every national election since Independence and respected the outcomes thereof further attests to the legitimacy that the regime enjoys. While the BJS and the BJP never opposed democracy, and as such cannot be considered anti-system parties even though they have been located on the far right of the ideological spectrum, the Communist parties have displayed rigid dogmatic adherence to the tenets of Marxism-Leninism, including their rejection of democracy, until recently. However, even  166  the CPI(M), the more radical of the two Communist parties, has recently revised its position when its government in West Bengal adopted an industrial policy that acknowledged the primacy of the private sector in promoting economic growth, welcomed foreign investment and called on workers to consider the effect that protecting theirrightsmay have on productivity and growth.292 2  FIGURE 4.5: V O T E R T U R N O U T IN N A T I O N A L ELECTIONS, 1952-1996  58  40 30 20 10 1952  1957  1962  1967  Source: Buder et al., India Decidesp.  1971  1977  1980  1984  1989  1991  1996  9; and Election Commission of India, GE96-LS19.  Although the legitimacy of the political system is high in India, the opposition's participation in consensus building has been virtually non-existent. The domination of the national political system for much of the country's post-Independence history by the Congress Party has prevented the opposition from being effectively and meaningfully involved in building a national consensus. As Rajni Kothari succinctly pointed out,  The role of the dominant party has been to evolve a consensus on both normative and procedural matters as well as major policy issues. In such a system, the dominant party becomes a norm setter for