ISSN (Print) - 0012-9976 | ISSN (Online) - 2349-8846
-A A +A

Caste and Electoral Outcomes

Misreading Hierarchy and the Illusion of Numbers

Dipankar Gupta ( taught sociology for nearly three decades in Jawaharlal Nehru University, New Delhi.

Understanding the relation of caste and electoral outcomes merely in terms of arithmetic runs is fundamentally fallacious. It fails to factor in the element of mutual repulsion among castes and the multiplicity of hierarchies. Shift from caste as a system to caste as an identity makes caste-arithmetic explanations of election results all the more questionable.

India has 900 million voters and yet it has successfully pulled off elections for seven decades. The world commends India as it should, for conducting this massive event in a generally peaceful and orderly fashion. Routinely also, scholars and pundits lick their pencils and dash off pieces on how castes impact voting behaviour.

India is probably the most stratified society in the world and, to top it all, it also has caste, that ultimate social curio, which no society anywhere else has. This is what makes otherwise cautious people become captives of an exotic frame of mind. It is this that distorts our understanding of the relationship between caste and politics.

Fundamental Fallacies

Narendra Modi’s unequivocal victory in 2019 is actually a repeat of his 2017 Uttar Pradesh (UP) performance (more than the 2014 general elections). On both occasions, Modi changed the rules of politics and let other parties pander to caste arithmetic. This identified them with specific groups, while Modi aimed for a consolidation of the rest. His opponents relied on the popular wisdom that caste numbers, if aggregated properly, hold the key to winning votes.

Such strategists were wrong in the past, but nobody paid too much attention to their mistakes as there were too many intervening variables that blocked one’s view. From 2014, if not earlier, it was becoming clear that correlating caste with electoral outcomes was problematic. Sadly, this realisation consistently appeared after the votes had been counted, not before.

One wonders why election experts are so wedded to an approach, a thought system, a methodology, even though they have been proved wrong time and again. If there is an overarching reason why this should be the case, it is that the overwhelming majority of scholars and election gurus have an elitist mindset. This needs to be kept in mind as we proceed with our main arguments.

Every time the election bugle blows, the tendency to rework caste arithmetic surfaces. If the numbers did not add up the right way in the past, it was because of a calculating error, and nothing more. What is hard to accept is that such analyses go wrong not because of something akin to a clerical error. The truth is that they are wrong because they rest on two fundamental propositions that are flawed to the core, but have escaped cross examination.

Error 1: The fallacy of proximate inter-caste solidarity: The first error lies in the belief that the castes that occupy roughly the same status in the Brahminical hierarchy, are natural allies, or can be tuned that way, with a little persuasion. This led them to conclude that rural peasants and artisans spontaneously club together. In other words, by virtue of belonging to a peasant caste, say Gujjar or Jat, a person is culturally conditioned to promoting fraternal ties with other peasant castes. The same logic was stretched upwards and sideways to cover the caste spectrum.

It was, therefore, believed that artisan castes, such as weavers and smiths, spontaneously bonded, as did prosperous communities like Thakurs and Baniyas and the various Scheduled Castes (SCs) too. Electoral wisdom, therefore, lay in banking on caste ideologies to charge political partisanship. In this, the reigning paradigm was that in Hindu India, people spontaneously accrete along hierarchically proximate caste blocs. If this is true in most aspects of their everyday life, why should this not hold for elections as well?

Error 2: The fallacy of caste numbers: The second mistake, which often stands alone, and is much more stoutly pronounced, is the assumption that certain castes numerically dominate certain parliamentary constituencies. This belief leads to the conclusion that there are several constituencies where a caste, such as the Yadavs, or Jats, or Kurmis, outnumbers all others. Naturally, the wisest thing to do, under these circumstances, is to woo these castes and win these seats.

As a consequence, caste leaders emerge as focal points for different political parties and a lot of energy is spent on dangling the appropriate carrot before them. These personalities begin to stand in for entire castes, as if on their beck and call the whole community would get up and march in formation. The only complication in all of this is when designated caste leaders compete among themselves. This then becomes a factional war and does not really disturb the principle of castes herding behind a supreme leader who is from among their own.

It is because most election experts operate with these two fallacies, indeed flaunt them, that their analyses and predictions are nearly always wrong. So, what is the truth and are we ready to face it?

Elitist Views of the ‘Other’

Before we start this exercise, we must exorcise the elitist, actually orientalist, bent of mind that most commentators and election specialists suffer from. For them, the Indian voter belongs to that category called the “other” who thinks differently, indeed exotically. The majority of them are Hindus first and everything else, including citizens, next. These “others” obey the primitive tug of primordial ideologies and place them above all other considerations. As long as this prejudice reigns among intellectuals, it is legitimate to ask: “Are we ready to face the truth?”

Predictably, as the general elections loomed in 2019, the familiar caste arguments were pulled out again. For example, TV broadcasters went on and on, everyday, about caste arithmetic and numbers manipulation. Now, post the election results, most of these channels do not say a word about caste arithmetic, for it failed comprehensively. India Today has, subsequently, come out with a research graphic showing that Other Backward Classes (OBCs) and Forward Castes (who should be on opposite sides) voted overwhelmingly for the Bharatiya Janata Party (BJP) and so did the majority of SCs. This made a mockery of TV jounalists’ earlier obsession with caste arithmetic and the supposed caste ideology that unites proximate castes. However, judging from past performance, it is quite likely that when the next elections come up, so will the familiar caste tales.

What needs to be underlined is that it is not because analysts calculated poll numbers incorrectly that their predictions went wrong. They were off target because they were bearing the burden of the two fallacies mentioned at the start, namely, the fallacy of spontaneous solidarity among proximate castes and the fallacy of numbers.

‘Mutual Repulsion’

The view that castes with similar lifestyles and occupations and who share the same berth in the Brahminical hierarchy form seamless solidarities is wrong. The major reason for this is that the Brahminical hierarchy is not accepted across the board by most other castes. To appreciate this it is necessary to realise that the fundamental property of caste, as Celestine Bougle (1991: 65) once elegantly put it, is that of “mutual repulsion.”

It is, therefore, incorrect to assume that certain castes naturally combine. Even when their occupations are the same, as between the peasant Jats and Gujjars, they see themselves as very distinct communities. They both may have earth in their fingernails, but that is where the likeness ends. They patriotically, and dogmatically, hold fast to their specific heritage, practices and beliefs (Gupta 1984), and look down upon all others, including other tillers of the soil.

What may appear as minor differences assume grand symbolic proportions in separating castes of similar occupational and economic status from one another. Castes semaphore differences on multiple axes. This task is never done parsimoniously, but with abandon, to the point of redundancy. For example, different fisherfolk castes draw boundaries among themselves on the basis of the sequence with which they weave their nets; from left to right, or in the reverse direction. Likewise, oil presser castes may distinguish themselves from another on the basis of whether the oil seeds are sun-roasted, or not, before they are crushed (Gupta 1984).

This is why the term “mutual repulsion” is so apt in the context of caste. Urban analysts, both national and international, on account of years of acculturation in a different universe, tend to bunch very different castes together. They then end up in believing that “forward” castes, such as the traditionally prosperous Brahmins and Thakurs, or SCs (ex-untouchables), such as Jatavs and Valmikis are monolithic election blocs. This ignores the deep cultural antipathy between castes in general. Much of this is fuelled by their respective origin tales and practices which are not just incommensurable, but competitive too.

We must then learn to appreciate another truth that our textbooks do not wisen us to. The sacerdotal renditions of caste, such as in the Purusha Sukta legend (attached to the Brahminical Rig Veda), do not fully reflect caste reality, and in particular, caste beliefs. According to this well-known version, Brahmins are on top followed by Kshatriyas (warriors), Vaishyas (peasants and merchants) and Shudras (servants and helots). But this is one myth among many.

There are literally hundreds of jati puranas,1 or caste origin tales, and in each one of them, the caste that forwards a particular myth places itself on top. This is equally true of jati puranas of SCs, as it is of other disparate warrior castes such as Gujjars, Sainis, Marawas, and Thevars. Even the Baniyas (merchant castes in the Vaishya bracket), though content with their status, nevertheless believe many illustrious rulers in history came from their community. What is more, some important merchant castes like the Khandelwals and Maheshwaris, abhor Brahmins for their alleged complicity in animal sacrifices (Babb 1998: 394–401).

There is, therefore, no jati (the endogamous, operative caste unit) which concedes that the bodily substances of its members are inferior to that of any other person. In which case, any question of a caste accepting that it is fundamentally impure or “polluting”—in terms of physical essences—is out of the question. In other words, there are many hierarchies, and it all depends on whose account you are listening to. If these other origin tales escape our attention it is because most of them were orally transmitted, unlike the Purusha Sukta story.

All castes are quite happy to declare their own superiority over the rest. Each of them clearly has a custom-made mirror on the wall. At this point, it must be added that though members of certain jatis, or castes, deny their status as “untouchables,” they are not necessarily against untouchability as such. While they powerfully resist being placed in that bracket, they might well believe that there are some other castes that deserve to be there.

Not surprising then that many jatis, widely considered to be “low,” have other ideas. They often claim, through their origin myths, a Kshatriya or Brahmin status, even though prosperous and established castes label them as despicable and defiling. Then there are jatis whose origin tales assert that they are even superior to Brahmins and Kshatriyas. According to their belief system, they were born of the gods (Shiva and Parvati, usually), or are offsprings of the sun and the moon, and this puts them way above other castes. If, after all this, they are today in an unenviable situation it is because of lost wars, or plain deceit, and not because they intrinsically deserve to be placed below others (Gupta 1984).

Such claims are really endless. There are thousands of jati puranas equalling probably the number of castes. The big truth, therefore, is that there exist many hierarchies in the caste order and not one, as most scholars and commentators tend to believe and propagate. Today, as more and more poorer caste folks are becoming literate, these tales, which were earlier orally transmitted, are now in a written form. Such assertions of caste dignity are coming up from all quarters with increasing frequency in contemporary times. This is because the caste system, as was ordained in tradition, is no longer effective and to understand this it is necessary to be mindful of the changes in rural India today. More of that in a while.

The primary truth we must come away with, after what we have just expressed, is the significance of “mutual repulsion” that governs inter-caste behaviour. In which case, when castes come together in political mobilisations, it is not caste ideology that does it but something else that lies outside, like urban jobs, or rural markets.

Misunderstood Demographics

Now we are ready to shift gears and come down to actual demographics. As a first step, we need to superimpose the caste presentation in the 1931 Census (the only place we have jati-wise numbers) on the political constituency map of India today.2 Once we do this, it will become clear that no caste has a numerical majority anywhere. In the best case scenario, the Marathas are about 32% of the population in Maharashtra. But this fact is neutralised because they are divided across the political bandwidth, from Congress to Shiv Sena, to BJP, and even to the Maoists.

This may come as a surprise but Yadavs make only about 7% of UP’s population while in west UP, the Jats are but about 8% in districts like Meerut and Muzaffarnagar. The reason why these two castes have been singled out is because much election writing talks about the dominant roles these communities play in vote capturing.

There are frequent statements that demarcate certain areas as Jat land and Yadav land. In the former, the owner-cultivator peasants are said to dominate, while in the latter, the traditional cowherd community is believed to be the most powerful. In Bihar’s Madhepura constituency, where Lalu Prasad Yadav and Sharad Yadav locked horns in an epic battle, the Yadavs were only about 23% of the population. In most other constituencies which are considered to be Yadav dominated, their numbers are only between 15% and 20%, and in none of them can you predict the poll outcome.

In Chhapra (Bihar), where the Yadavs number between 15% and 20%, in the heyday of caste politics in the 1990s, the honours were shared between Rashtriya Janata Dal, BJP and Janata Dal (United). In places like Balia, Motihari and Madhubani, in Bihar, where the caste distribution is similar, election results vary a lot. In west UP, the supposed Jat stronghold, the BJP was an outlier for 16 years till 2014 when it trumped all other parties. In that period in between, Congress, Samajwadi Party (moniker, Yadav party) and Bahujan Samaj Party (BSP) (moniker, Scheduled Caste party) came to power in succession. In the late 1960s and early 1970s, even the Communist Party won in a west UP constituency, but the Rashtriya Lok Dal (moniker, Jat party) never made it ever. In Bijnor, another supposed Jat stronghold, once again all-India parties, like the Congress party and the BJP, won in recent elections. The Jat party, true to past form, was nowhere in sight. There is no need to further belabour this point.

If one were to look at the way SCs vote, once again there is no consistency. For example, in areas like Sonbhadra and Lalitpur (in UP), where BSP lost in 2002, it won convincingly in 2007. In Pilibhit and Bareilly where SC numbers are low, the BSP won in 2007. In 2002, however, the BSP lost in Sonbhadra even though this constituency has a high SC population. Interestingly, the BSP won every seat in Sonbhadra and Lalitpur in 2007, but lost in these areas in 2002. In Hardoi, the SC population is high but the BSP lost a bulk of the seats both in 2002 and 2007. Nor can we correlate BSP’s performance with SC literacy rates (Gupta 2010: 284–85).

Therefore, to assume that Jats or Yadavs or Jatavs, or any other caste, call the shots in parliamentary elections is false. This optical illusion came about because Yadavs in east UP, or Jats in west UP, were at one time the best endowed in terms of cultural, social, and good old- fashioned capital. Matters changed quite significantly from the 1990s. In time, erstwhile supplicant castes, buoyed by urbanisation and reservations, could parade their own band of literati and did not need to rely on Jat and Yadav virtuosos.

With this transformation, certain constituencies that earlier sported the tag of being Jat-, Gujjar-, or Jatav-dominated lost their unique distinction. As other peasant, artisan and Dalit castes now had their own cultural specialists, including government officials, they could amplify their aspirations independently. They did not need the help of Jats or Gujjars and, therefore, lost nothing in mediation. This, as we have already mentioned, led to a proliferation of written jati origin tales which were earlier only orally transmitted. In addition, the tag of a constituency being a Jat or Gujjar, or whatever, looked less convincing.

To reiterate, it is hard to predict election outcomes on the basis of numbers for two reasons. The first is that there are always good non-caste (or “secular”) considerations why people of the same caste vote differently. Second, in any constituency there would typically be about five to six castes of roughly equal strength. Now, as there are generally two main contenders in most electoral contests, and as nobody wants to waste a vote, most people must, by necessity, cast their ballot in favour of somebody outside their caste.

Like it or not, this is a logical outcome. Regardless of how wedded one might be to one’s caste identity, at parliamentary election time the majority are compelled to vote outside their own caste. It is also wise to remember that for Narendra Modi, in 2017, to get about 40% of votes in UP, many hitherto antagonistic castes must have jointly contributed to this outcome. No one caste would have succeeded in bringing this about; they just do not have the numbers. The India Today statistic, referred to earlier, which says that an equal number of OBCs and Forward Castes voted for BJP demonstrates this point.

It is also clear that caste ideologies could not have prompted this merger either (refer to “Fallacy number 1” above), therefore, something else did. But we will never begin to fathom that one out as long as we are looking specifically for caste correlations. Do not forget either, that people are often not aware of all those others with whom they are supposed to share space in the rubrics that are designed for political purposes.

For example, Kurmis, Nonias, and Koeris of east UP officially belong to the OBC category, but they may not know of one another’s existence. This is because their social lives do not overlap, even though they are not separated by wide distances. Yet, we urban experts have no hesitation to group them together as OBCs and treat them as a bloc just because they look, dress and talk in a very similar fashion. That, as should be easy to persuade readers by now, is a superficial reality. Given this truth, how then can you expect caste coalitions to jump mud walls and link across whole states, let alone between states, on the strength of primordial loyalties alone?

Caste Identity

The caste order has over the past six decades or so, gradually metamorphosed and this transformation is sociologically very significant. While there were always disputes about who occupied the top spot in a pan-Hindu setting, at the local level (which mattered the most) this issue was easily resolved. The rural oligarchs of the region, those with superior economic and political power, positioned themselves at the apex of the caste pyramid.

In actual terms, regardless of what Manu (or any other ancient Brahmin sage) may have said, the Brahmin religious functionary was placed below the leading feudal family. The landed, rural elite, accorded Brahmin priests a higher status than those of other service castes, but refused to acknowledge them as their equals (Quigley 1993: 54–60; Raheja 1988: 34). For instance, should an elderly Brahmin in Rajasthan come in the presence of a young Rajput landlord, the Brahmin is supposed to bow low in salutation. This practice is no longer followed quite that punctiliously, but it exists in living memory even today.

The dismantling of the old feudal elite began with the abolition of “zamindari,” or landlordism, and this process started almost as soon as India became independent. Over the past 70 years, with each passing generation, landholdings kept getting subdivided. This reduced a majority of old landed families to the status of small and middle farmers. The smarter ones among the then rural rich, sold their land, fattened their wallets and left the village in the 1960s and 1970s.

Those who stayed back gradually wasted away and saw their power move to the hands of the more numerous peasant communities now that every vote had equal weight. They clung to memories of the days when their grandparents mattered, but lacked the resources to be remotely like them. With the breakdown of the traditional rural economy, it was not just the patron–client networks that died, but so did the long established caste hierarchy of the region.

Poor peasant and other service castes could now vocally express their caste pride in a way they would have never dared to do in the past. Interestingly, even till the 1950s, these other subordinate castes hesitated to call themselves Kshatriyas of any kind as that would make their local overlords very angry. The Kshatriya status was reserved for the highest in the realm and lesser communities dared not appropriate it.

It was, in fact, safer in those days to ascribe oneself as a Brahmin for, in the village context, then and now, a Brahmin has little power. Though this is contradictory to the traditional Brahminical script, no landlord would swap Kshatriya status for a Brahmin one. So, if some so-called “lower” castes had Brahminical aspirations, it would be indulgently smiled over by the warrior landed castes. So long as these indigent service castes steered clear of Kshatriya ambitions, everyone stayed well.

Even where Brahmins are traditionally landlords, should a down-and-out caste announce a priestly status, it would probably go unnoticed at the upper levels. The Jhas of Bihar and the Anavils of Gujarat are Brahmins, but they see themselves as landed elite first. They do not consider the Brahmin priest, even the most exalted one, equal to them in status and would never agree to kinship ties with them.

Hostilities towards so-called lower castes claiming Brahmin, Kshatriya, or even semi-divine status are, more or less, a thing of the past. It rarely occurs today, and when it does, it immediately attracts attention and is not par for the course as in earlier times. This is largely on account of the fact that the caste system has collapsed as the caste hierarchy can no longer be enforced by the local feudal elite.

That does not imply the end of caste. Instead, what we see is a resurgence of caste identity like never before. As the caste system is now far from being robust, it is easier for subaltern castes to loudly declare their long held elevated self-image; an act they would never have contemplated in earlier epochs. They may have subscribed to this imagined status all along but as the closed agrarian economy held, with the rural overlord on top, the caste system of the day was not to be trifled with.

Now, as all that has changed, they can extravert their inner sentiments, and origin tales, without fear of reprisal. What was once, at best, said sotto voce, is now declared in a full-throated fashion. In the past, when the caste system reigned, any such open defiance of the locally established hierarchy would be met with physical punishment, including death. Not quite that way, anymore. The caste system is dead, long live caste identity. This is why it makes better sense to look at the castes today as horizontal competing entities and not as obedient servants of an established ritual hierarchy.

Nor do old status markers and sartorial restrictions matter like they once did. Less privileged castes now indulgently twirl their moustaches, wear elaborate headgear, and even ride a horse when they go to the bride’s home to get married. These were considered effronteries by the old elite till a few decades back, but that was when the caste system was doing well, as a system. If they find an alternative source of employment, members of service castes, such as scavengers, drum beaters, and palanquin bearers, move on. They are reluctant to perform their traditional functions for they see them as degrading. The closed and encased landed economy would have earlier prevented them from exercising this option, but that is soon becoming history.

The opening up of the rural economy has many consequences, but the most important one is that it has generated newer occupational opportunities. The poor may not have become much richer, but they are poor differently today because they are untied. This should also be linked to the atrophying of jajmani relations where payments were made in kind and settled by tradition. Contract relations are now becoming the order of the day (Breman 1974; Karanth 2004).

What Makes Castes Unite?

When we appreciate the two overwhelming truths, namely that castes mutually repel each other and that demographically no caste has numerical dominance by far, certain issues suddenly become clearer. Today, in the caste order, where identities reign, there are no ideological friends, only enemies and rivals. In which case, when different castes come together for political reasons, it is not caste logic that is drawing them close, but something outside of it. This calls for greater sensitisation to inputs from the secularised economic and social world in caste analyses, else we will keep missing the point. It is also impossible not to notice the many differences that exist between members of a single caste and how that leads to diverse political affiliations.

The coming together of the OBCs illustrates this point rather well. In 1991, India granted “reservations” (a quota-based affirmative action system) to a number of small and medium peasant castes as they were considered socially and educationally backward. It was deemed proper, and democratically correct, that the state help these communities to move up in society and avail of urban economic opportunities. To this end, quotas were established which guaranteed a certain percentage of seats to them in government educational institutions and jobs.

Thus, when these disparate rural castes came together they had a single goal in mind, namely economic betterment. They were not motivated by caste ideology as each had specific origin tales and observed their very own social/ritual practices. Now that OBC reservations are well in place there are tensions between those who once functioned as a pack, on who gets how much of the quota pie. Also, over time, other political questions have arisen pushing the settled issue of OBC reservations to the back. As a result, OBCs, that earlier glued together, are now in political and economic combat. Nor can any one caste be assumed to be the winner as the numbers are not there on anybody’s side. Nor, to be boringly repetitive, caste ideologies by themselves could have prompted this merger either.

In conclusion, it is disturbing when commentators spontaneously tend to view election results primarily through the caste optic. Elementary statistics will tell you that it is not difficult to muster up some kind of correlation between any two, or more, categories. If, instead of castes, election experts were to correlate first generation urbanites, out of work weavers, vocationally qualified workers, ex-landlords, agricultural sharecroppers, petty bureaucrats, and so on, with election results, another set of correlations would emerge.

Given what has just been said about caste numbers, and the problematic nature of hierarchical ranking, any easy correlations between caste and election outcomes should be questioned. In fact, if one sees such a correlation then it has to be explained in terms of other variables and not considered to be a foregone, unproblematic, conclusion.


1 Jati is the local term signifying the operative, endogamous, caste unit.

2 After 1931, the colonial government did not conduct a census in 1941 for it was finding it difficult to finance this project given the huge expenses that World War II had imposed on Britain. Post independence, the newly formed Government of India resumed the census but kept out actual jati enumerations for it believed that this would encourage caste politics.


Babb, L A (1998): “Rejecting Violence: Sacrifice and Social Identity of Trading Communities,” Contributions to Indian Sociology (NS), Vol 32, pp 387–407.

Bougle, Celesin (1991): “The Essence and Reality of the Caste System,” Social Stratification, Dipankar Gupta (ed), Delhi: Oxford University Press, pp 64–73.

Breman, Jan (1974): Patronage and Exploitation: Changing Agrarian Relations in South Gujarat, India, Berkeley: University of California Press.

Gupta, Dipankar (1984): “Continuous Hierarchies and Discrete Castes,” Economic & Political Weekly, Vol 19, Nos 46–48.

— (2000): Interrogating Caste: Understanding Hierarchy and Difference in Indian Society, New Delhi: Penguin.

Gupta, Dipankar and Yogesh Kumar (2007): “When the Caste Calculus Fails: Analysing BSP’s Victory in UP,” Economic & Political Weekly, Vol 42, No 33, see also Gupta, Dipankar, 2010, The Caged Phoenix: Can India Fly?, Stanford: Stanford University Press.

Karanth G K (2004): “Replication or Dissent? Culture and Institution among ‘Untouchable’” Scheduled Castes of India,” Caste in Question: Identity or Hierarchy, Dipankar Gupta (ed), New Delhi: Sage, pp 137–64.

Quigley, Declan (1993): The Interpretation of Caste, Oxford: Clarendon Press.

Raheja, Gloria G (1988): The Poison in the Gift: Ritual, Prestation and the Dominant Caste in a North Indian Village, Chicago: University of Chicago Press.

Updated On : 26th Jun, 2019


(-) Hide

EPW looks forward to your comments. Please note that comments are moderated as per our comments policy. They may take some time to appear. A comment, if suitable, may be selected for publication in the Letters pages of EPW.

Back to Top