This chapter builds upon the link between limited information and ethnic categorization to develop a theory of individual decision making among benefit-seeking voters and office-seeking elites in patronage-democracies. Confronted with competing elites, how does a voter in a patronagedemocracy decide who is most likely to channel benefits to him individually? And confronted with different strategies for the distribution of benefits, how does an individual elite seeking to build a following decide whom to target? The voting decision in patronage-democracies, I argue, typically approximates a situation in which observers (voters) are forced to distinguish between individuals (the recipients of past patronage transactions) under severe information constraints. These severe information constraints produce a self-enforcing and reinforcing equilibrium of ethnic favouritism: voters expect co-ethnic elites to favour them in the distribution of benefits, and elites expect co-ethnic voters to favour them in the distribution of votes (see Figure 1.1, p. 12).

The tendency of patronage politics to go hand in hand with expectations of ethnic favouritism has been noted in other theoretical and empirical work. According to Kearney, a student of Sri Lanka: “A common expectation seems to be that a person holding a public office or other position of power will use his position for the near-exclusive benefit of his ‘own’ people, defined by kinship, caste, ethnic community or personal loyalty.” According to Haroun Adamu, a student of Nigerian politics: “It is strongly believed in this country that if you do not have one of your own kind in the local, state and/or national decision-making bodies, nobody would care to take your troubles before the decision makers, much less find solutions to them.” Kenneth Post's description of elections in Nigeria emphasizes much the same point:

This chapter describes the formation of the BSP in Uttar Pradesh, Punjab, and Karnataka. In Part I of this book, I argued that upwardly mobile elites in patronage-democracies seek a channel to political office in the long term but are indifferent regarding the type of party – ethnic, multiethnic, or nonethnic – that offers them such a channel. Here, I describe how, in the states of Uttar Pradesh and Punjab, a representational blockage in the competition diverted these elites toward the BSP and so enabled it to acquire the critical mass of support from elite activists necessary for party formation. In Karnataka, however, representational “openness” in the competition for these Scheduled Caste elites prevented the BSP from obtaining their support. As a consequence, the party has not been able to present itself as a viable option on the electoral market in Karnataka, despite repeated attempts.

The chapter is organized as follows: Section I describes the emergence of a class of upwardly mobile, office-seeking elites among Scheduled Castes in each of the three states under study. Section II describes the variation in opportunities for representation available to these elites in the major parties in each of these states. It shows that the Congress party and the rest of the party system in Uttar Pradesh and Punjab blocked the entry of these Scheduled Caste elites. In Karnataka, however, the Congress party and other political parties were able to incorporate these elites successfully. Section III describes the link between the representational blockage of Scheduled Caste elites and the formation of the BSP in Uttar Pradesh; Section IV does the same for Punjab; and Section V describes the link between representational “openness” and the failure of the BSP to take root as a viable political party in Karnataka.

The Emergence of Office–seeking Elites among Scheduled Castes

At independence from the British in 1947, Scheduled Castes in India participated in electoral politics mainly as benefit seekers. Less than 10 percent of the Scheduled Caste population was literate, whereas the literacy rate among the general population was at least twice as high.

In Part I, I argued that the patronage-democracy was a distinct type of democracy, with distinct patterns of voting behaviour and therefore distinct conditions leading to ethnic party success. In this chapter, I propose an interpretation of the Indian political system as a patronage-democracy. This interpretation applies broadly to postcolonial India, from independence in 1947 until the present. The Indian state is currently undergoing two major structural changes: the deregulation of the Indian economy and the decentralization of political power. So far, however, these changes have not been of sufficient magnitude to alter the nature of India's patronagedemocracy.

Virtually every major study of Indian politics and political economy has remarked upon India's dominant state and the extent to which voting behaviour in India is conditioned by expectations of access to the state.4 This chapter builds upon this previous literature in order to advance two further arguments. First, it argues that the essential element influencing voting behaviour in India is not simply the dominance of the state, but the ability of those who control the state to exercise discretion in the implementation of state policy. Second, whereas previous studies have analyzed state dominance and ethnic politics as two parallel processes, this chapter proposes a connection between the two: The individualized distribution of benefits at the disposal of those who control the dominant state, I argue here, produces and sustains ethnic politics in India.

Section I describes the discretionary control of the Indian state over jobs, livelihoods, and living conditions and shows that those who hold state office have considerable opportunity for rent seeking in the implementation of state policy.

Since census data in India are not available at the constituency level, it is not possible to obtain a precise measure of the urban population in a constituency. I classify a constituency as “urban” here if (1) the urban population in the district in which it is contained is higher than the average urban population of India as a whole (25.7 percent, according to the 1991 census) and (2) if it contains one or more towns in the district. Parliamentary constituencies can often span more than one district. However, those listed here are almost all completely contained within district boundaries, although the district itself may contain more than one parliamentary constituency. Most urban constituencies here are likely to have large (sometimes majority) rural populations. However, the urban clusters they contain make them different from constituencies that are entirely rural.

This chapter describes the Bahujan Samaj Party (BSP) and its target ethnic category, the Scheduled Castes. Section I provides a brief history of the BSP and locates it in the context of Indian politics. Section II discusses the Scheduled Castes. Section III describes the platform on which the BSP appeals to the Scheduled Castes and contrasts its platform with that of the Congress party, its principal competitor. Section IV describes the method that I employ to develop the hypothesis proposed in this book. Section V evaluates this hypothesis against the principal explanations for the variation in support for the BSP among Scheduled Castes proposed by informed observers of the BSP, including social scientists, media analysts, and BSP activists themselves. Subsequent chapters then develop and test the separate propositions contained in this argument.

History

The Bahujan Samaj Party was founded in 1984 by Kanshi Ram, a Ramdassia Sikh belonging to the Chamar caste category among the Scheduled Castes. Born in 1934 in a rural district in the north Indian state of Punjab, Kanshi Ram obtained a B.Sc. degree from a local college. He then entered one of the government positions “reserved” for Scheduled Caste candidates, first with the Survey of India and then with the government-owned Explosives Research and Development Laboratory in the city of Pune in the western state of Maharashtra.

Although he did not initially aspire to a political career, incidents of caste discrimination during his tenure in government service pushed him toward politics. While employed in the Pune Laboratory, Kanshi Ram became involved in an agitation by Scheduled Caste employees protesting the cancellation of an official holiday commemorating the birthday of Dr. B. R. Ambedkar, India's most prominent Scheduled Caste leader during the movement for independence from the British. He was suspended following an altercation with an upper-caste superior over this incident. He resigned from his job altogether in 1971 and, in 1973, founded the Backward and Minority Community Employees Federation (BAMCEF), a local association of government employees with the explicit purpose of increasing the political influence of the ethnic categories that they represented.

In its initial years, BAMCEF supported the activities of the Republican Party of India, then India's best-known Scheduled Caste party, with a base mainly in the state of Maharashtra.

What determines the initial differences in relative opportunities that a political party in a patronage-democracy offers to elites from different ethnic groups? And, regardless of these initial differences, once it is clear that success depends upon acquiring the correct head count, why might unsuccessful parties fail to adjust their ethnic profiles over repeated elections?

These questions raise a broader theoretical question: Under what conditions are political parties in multiethnic democracies able to incorporate new elites, and under what conditions are they likely to fail? General theories of party politics in multiethnic societies propose a sociological answer to the problem of elite incorporation, according to which success or failure in elite incorporation is determined by the pattern of conflict or harmony between ethnic categories in society. Where ethnic categories are in conflict in society, according to these theories, elites belonging to these categories will not be able to coexist in the same political party. Consequently, multiethnic parties are described as “inherently unstable.” Conversely, ethnic parties, which bring together elites from a single ethnic category, or from socially harmonious ethnic categories, are posited to enjoy a more stable existence.

This chapter proposes a model of elite incorporation that explains the successful incorporation of elites by any political party – whether multiethnic, nonethnic, or ethnic – as a consequence of the internal organizational structure of the party. The model builds upon Myron Weiner's study of the incorporation of new elites in the Congress party in India. According to the model, parties with competitive rules for intraparty advancement, other things being equal, are able to continually incorporate new elites while keeping old ones acquiescent. Such parties are stable parties, better able to retain the allegiance of elites during “lean” periods when they are out of government. Parties with centralized rules of intraparty advancement, other things being equal, are “closed” to the entry of new elites. Such parties are unstable, deeply vulnerable to defections by old elites when they are out of power. The internal organizational structure, I propose, is more influential than the degree of conflict between ethnic groups in society. Where a competitive structure exists, it promotes coexistence even between elites from warring groups in the same party. Where a centralized organizational structure exists, however, we are likely to see splits and defections even within a party dominated by elites from the same ethnic category.

This chapter applies the hypothesis developed through the study of the BSP to three other ethnic parties in India: the pro-Hindu Bharatiya Janata Party (BJP) in India's 1991 parliamentary elections (section I); the pro-Tamil Dravida Munnetra Kazagham(DMK)in the state of Tamil Nadu in the 1967 legislative assembly elections (section II); and the pro-Jharkhand Jharkhand Mukti Morcha (JMM), a regionalist party that called for the creation of a new state of Jharkhand out of the state of Bihar and its neighbours, in the 2000 legislative assembly elections in Bihar (section III). I rely for these analyses mainly on several excellent secondary studies of these parties and their vote bases. The data available from these studies are not sufficiently precise to permit a systematic evaluation of the argument. Instead, I hope to show only that in each case, the hypothesis proposed in this book generates a more plausible explanation than the alternatives for an otherwise puzzling outcome.

I select these three parties in order to maximize the range of variation in the data, keeping the condition of patronage-democracy constant. The three parties differ, first, in the type of category that they sought to mobilize. The BJP in 1991 targeted the religion-based category “Hindu.” TheDMK defined its target ethnic category in 1967 as a linguistic/regional category of Tamil speakers in Tamil Nadu. The Jharkhand Mukti Morcha defined its target ethnic category in 2000 as “Jharkhandis,” by which it meant all inhabitants of the Jharkhand region. Second, they differ in the relative size of their target ethnic categories: Hindus in India and Tamil speakers in Tamil Nadu are both majority categories, while Jharkhandis constituted a regionally concentrated minority within Bihar. Third, they differ in the size of the political arena in which they operated: The BJP in 1991 was a nationwide party; theDMKin 1967 concentrated its efforts in the southern state of Tamil Nadu; and the Jharkhand Mukti Morcha in 2000 contested in the state of Bihar. Finally, the three parties differ in the type of elections and the time period chosen for analysis.

This chapter builds upon the individual microfoundations introduced in the previous chapter to propose a hypothesis predicting ethnic party success and failure in patronage-democracies. Voters in patronage-democracies, I argue here, decide whether or not to vote for any type of party – ethnic, multiethnic, or nonethnic – based on a comprehensive process of ethnic head counting. They formulate preferences across parties by counting the heads of co-ethnics in positions of power and prestige in party organizations. They decide to vote in accordance with these preferences only when they expect their preferred party to win or to exercise influence after the election. And they formulate expectations about the probable electoral outcome by ascertaining, after counting the heads of co-ethnics and ethnic outsiders in the electorate, that members of their “own” ethnic category are numerous enough to take the party past the threshold of winning or influence. Ethnic parties are likely to succeed, therefore, only when they have a comparative advantage in the representation of elites from their target ethnic category and when members of their target ethnic category are numerous enough to affect the outcome. When either of these two conditions is not met, ethnic parties are likely to fail.

This chapter is organized as follows: Section I outlines the proposition that voters in patronage-democracies formulate preferences across political parties by counting heads belonging to their “own” ethnic category rather than by assessing party issue positions; section II outlines the proposition that these benefit-seeing voters are also strategic actors who will vote for their preferred party only if it is likely to win or to exercise influence after the election, and not otherwise; section III outlines the proposition that voters formulate expectations about the electoral outcome by counting the heads of co-ethnics across the electorate; section IV outlines the proposition that party strategy in patronage-democracies should give primary emphasis to developing advantageous systems of ethnic categorization rather than to developing advantageous issue dimensions; section V, based on these previous propositions, outlines the hypothesis predicting the success of ethnic parties; and Section VI elaborates on the paradox of ethnic party failure. The argument presented here also generates a range of additional hypotheses about elite strategy and voting behaviour in patronage-democracies. These are identified but not tested here, since they are tangential to the main subject of this study.

Why do ethnic parties succeed in obtaining the support of members of their target ethnic group(s)? ethnic political parties now flourish across the democratic world. canada, spain, india, the united kingdom, israel, sri lanka, macedonia, south africa, and russia are only a few examples of the established or emerging democracies in which they have taken root. for social scientists interested in explaining important political phenomena, the question is worth asking for its own sake. at the same time, the answer has broader implications for those with a stake in the survival of democratic regimes. ethnic parties, and the politicization of ethnic differences more generally, are presumed to constitute a major threat to democratic stability.1 an exploration of the processes by which such parties succeed or fail, then, illuminates also the processes that undermine or preserve democracy.

Drawing on a study of variation in the performance of ethnic parties in India, this book proposes a theory of ethnic party performance in one distinct family of democracies, identified here as “patronage-democracies.” Voters in patronage-democracies, I argue, choose between parties by conducting ethnic head counts rather than by comparing policy platforms or ideological positions. They formulate preferences across parties by counting the heads of co-ethnics across party personnel, preferring that party that provides greatest representation to their co-ethnics. They formulate expectations about the likely electoral outcome by counting the heads of co-ethnics across the electorate. And they vote for their preferred party only when their co-ethnics are sufficiently numerous to take it to a winning or influential position.

This process of ethnic head counting is the foundation for the central argument advanced in this book: An ethnic party is likely to succeed in a patronage-democracy when it has competitive rules for intraparty advancementandwhen the size of the ethnic group(s) it seeks to mobilize exceeds the threshold of winning or leverage imposed by the electoral system. Competitive rules for intraparty advancement, other things equal, give a party a comparative advantage in the representation of elites from its target ethnic category. And a positive difference between the size of its target ethnic category and the threshold of winning or leverage indicates that the party has a viable shot at victory or influence.

The previous chapter argued that a representational blockage in the competition was necessary in order to induce Scheduled Caste elites to join the BSP. In this chapter, I link representation with voter preferences. I show that Scheduled Caste voters in Uttar Pradesh and Punjab formulate preferences by counting heads across parties rather than by examining the issues that the parties stand on. Because the BSP had a monopoly on the representation of Scheduled Caste elites in these two states between 1984 and 1998, Scheduled Caste voters preferred the BSP to other parties. However, voters from the Scheduled Caste categories that were better represented in the BSP were more likely to prefer it than voters from underrepresented Scheduled Caste categories. The link between representation and voter preferences, furthermore, holds not just for Scheduled Caste voters but for voters from all other ethnic categories as well.

The data that I use to develop and test this proposition are drawn from across four elections in India, held between 1996 and 1998. For information about the ethnic profile of political parties, I rely on interviews with party personnel in each state conducted between 1996 and 1998. For party issue positions, I conduct a content analysis of campaign messages during the 1996 parliamentary election campaign. For data on perceptions of salient issues among Scheduled Caste voters, and on Scheduled Caste voter preferences independent of voting behaviour, I draw upon survey data from the Indian National Election Studies for the 1996 parliamentary election. Finally, for data on the voter preferences of other ethnic categories from other elections, I rely upon indirect evidence drawn from party and candidate strategies collected during four election campaigns.

The chapter is organized as follows: Section I predicts the preference distribution of Scheduled Caste voters across political parties in the states of Uttar Pradesh and Punjab if they formulate preferences by counting heads. (I exclude Karnataka from the analysis of Scheduled Caste voter preferences, since the sample of Scheduled Caste voters who expressed a preference is too small to be statistically reliable.) Section II predicts the preference distribution of Scheduled Caste voters across political parties if they formulate preferences by assessing party issue positions. Section III shows that the first prediction captures the distribution of Scheduled Caste voter preferences more accurately than the second. Section IV extends this proposition to voters from other ethnic categories in other states.

The purpose of this appendix is to provide the reader with sufficient information on the ecological inference (EI) method to evaluate the method with which I arrive at the estimates of Scheduled Caste support for the BSP (section I), to evaluate the use of these estimates for the second-stage analysis of strategic voting (Section II), and to underline the possibilities for research on ethnic politics created by EI (section III). It would be most useful for the reader if read in combination with King's A Solution to the Ecological Inference Problem (Princeton: Princeton University Press, 1997).

Use of EI to Generate Constituency-Level Estimates of Scheduled Caste Voting Behaviour

The ecological inference problem that I address here is how to estimate the proportion of Scheduled Castes who voted for the BSP in each constituency. King's method solves this problem in two iterations. In the first iteration, it is used to estimate the proportion of Scheduled Castes who turned out in each constituency. The logic of the method is as follows. For any constituency, we can compute deterministic bounds on turnout rates for Scheduled Castes and non–Scheduled Castes in the population. Once we have these deterministic bounds for each constituency, EI computes a probabilistic model showing where the turnout rates are likely to lie by “borrowing strength” from all constituencies in order to produce estimates for each. Using King's notation, the ecological inference problem at iteration 1 is as follows:

Xi = Proportion of SCs in the electorate in constituency i

1 − Xi = Proportion of non-SCs in the electorate in constituency i

Ti = Proportion of electors who turn out to vote in constituency i

1 − Ti = Proportion of electors who do not turn out in constituency i

βib = Proportion of Scheduled Castes who turn out in constituency i

1 − βib = Proportion of Scheduled Castes who do not turn out in constituency i

βiw = Proportion of non–Scheduled Castes who turn out in constituency i

1 − βiw = Proportion of non–Scheduled Castes who do not turn out in constituency i