An analysis by CSDS based on the electoral outcome and the findings of a nation-wide post-election survey.
YOGENDRA YADAV with SANJAY KUMAR and OLIVER HEATH
THE results of the 13th Lok Sabha elections mark the completion of a process that has brought about the Bharatiya Janata Party's rise to power, a process that began with the 10th Lok Sabha elections in 1989 and has continued uninterrupted since.
As was argued in the first part of this analysis (Frontline, November 5), the results of the recent elections do not point to a clear victory for the BJP, or for that matter for any other national party. Yet in more than one sense the BJP has completed one leg of a remarkable political journey. Its performance in Goa, Assam and a few other States at the geographical periphery symbolises the geographical expansion that the BJP has gone through; at the end of this round of elections, there are very few States - Kerala, for instance - where the BJP or at least one of its allies is not one of the main competitors in the electoral arena. Its alliance with the Dravida Munnetra Kazhagam (DMK) in Tamil Nadu completed the process of its political expansion from a party that was once considered a political "untouchable" to one that is acceptable to everyone across the ideological spectrum, except the Congress(I) and the Left parties.
In this series, the same group of electors, comprising a national representative sample, were approached after each of the last three general elections and face-to-face interviews conducted to find out their voting behaviour and political opinions and attitudes. Retaining the same sample, after making minor adjustments to accommodate new voters and factors such as migration, makes it possible to study change over time by comparing the results.
The survey was carried out in 105 parliamentary constituencies - it covered 419 polling stations in 210 Assembly segments in 20 States and Union Territories. A total of 9,111 electors were interviewed in their homes in the week after each phase of polling. The last round of the survey was completed on October 5, the day before the counting began. The constituency-wise results of the Lok Sabha and Assembly elections have been taken from the Web site of the Election Commission and recomputed by the CSDS data unit.
In order to promote greater transparency in survey reporting, the CSDS and Frontline would like to share with readers some background information about the survey.
The social composition of the respondents in this survey is as follows: 49.3 per cent women, 18.3 per cent Dalits, 7.9 per cent Adivasis, 10.3 per cent Muslims, 77.8 per cent rural voters, 41 per cent uneducated voters and 6.7 per cent graduates. There is thus an over-representation of rural voters and an under-representation of Muslim voters. In terms of reported voting, the sample percentages (actual results in brackets) are: the BJP and its allies 38.8 (40.8), the Congress(I) and its allies 36.7 (33.8), the Left 7.6 (8.0) and the Bahujan Samaj Party 3.1 (4.2). The sample size for each State and significant discrepancies, if any, have been mentioned in the analysis for each State.
The survey was designed and coordinated at the national level by Prof. V.B. Singh (Principal Coordinator), Sanjay Kumar (National Coordinator) and Yogendra Yadav, all from the CSDS. A network of scholars worked with the CSDS in designing and executing the survey in different parts of the country: Dr. K.C. Suri (Andhra Pradesh), Prof A.K. Baruah and Dr. Sandhya Goswami (Assam and other Northeastern States), Rajendra Ravi (Bihar), Prof. Peter deSouza (Goa), Dr. P.M. Patel (Gujarat), Dr. Jitendra Prasad (Haryana), Prof. T.R. Sharma (Himachal Pradesh and Punjab), Dr. Sandeep Shastri (Karnataka), Prof. G. Gopa Kumar (Kerala), Dr. Ram Shankar (Madhya Pradesh), Prof. Suhas Palshikar (Maharashtra), Dr. S. N. Misra (Orissa), Dr. Sanjay Lodha (Rajasthan), Prof. G. Koteshwar Prasad (Tamil Nadu and Pondicherry), Dr. Pradeep Kumar, V.K. Rai and Dr. A.K. Verma (Uttar Pradesh) and Prof. A.K. Chaudhuri (West Bengal). The members of the central team at the CSDS are: Himanshu Bhattacharya, Mona Gupta, Oliver Heath, Sudhir Hilsayan, K.A.Q.A. Hilal, Bhaskar Jha, Angad Kumar, Kanchan Malhotra, Anindya Saha and Chitrali Singh.
Both these expansions were preconditions for the third process of social expansion: the BJP's changeover from being seen as a party of the urbanised and of the Bania-Brahmin to becoming a party with a wider social base. The journey was not without its share of problems and retreats: its uneven performance in this election testifies to that fact. More important, the BJP did not carry out this journey on its own terms. Quite early on, it realised that it was on a new turf. It was forced to come to terms and negotiate with a regionalised polity and the democratic upsurge of the "shudras". That negotiation process is far from over and its outcome is still an open question. But the BJP's arrival as a party that today heads a coalition that is seen to be less fragile than at least four previous governments perhaps occasions stock-taking to get a long-term view of what this journey has meant for India's democratic polity.
The BJP's rise to power is related to a fundamental process of change that goes beyond routine electoral victories and defeats and the mechanics of government formation and alteration. In its multi-pronged attempts to win an electoral majority and create for itself a somewhat stable support base, the BJP has created a new social bloc, a new coalition of various social groups, that now lays claims to political power. The BJP's attempt to become the 'natural party of governance' has resulted in a new kind of majoritarianism, which is deeper and subtler than what both its spokespersons and its critics would allow for. It is not a simple Hindu majoritarianism. Although religious symbolism has been a trademark of the BJP's mobilisational strategy, and religious exclusions continue to mark the boundaries of the new social constituency, religion is not one of the principal faultlines in the creation of this new social bloc. The new social bloc is formed by the convergence of traditional caste-community differences and class distinctions. It may be an exaggeration to say that the BJP represents the rebellion of the elite, but it is nevertheless true that its rise to political power has been accompanied by the emergence of a new social group that is defined by an overlap of social and economic privileges.
It is important to interpret the signals of the 1999 verdict in the light of this process. For an obsession with the game of numbers or the business of alleged vote banks can draw attention away from the larger picture. Since the official data on the elections are not very helpful in learning about the social basis of the voters' choices - of who voted for whom - we shall mainly use here the post-poll survey conducted by the Centre for the Study of Developing Societies and co-sponsored by Frontline. This survey is part of a longer series of National Election Studies conducted by the CSDS and is one of the few sources of reliable data with a social science perspective and which allows one to see the current picture in a historical perspective.
BEFORE one looks at the precise shape and nature of the new social bloc, let us first take a look at the overall picture of flow of votes between the 1998 elections and the latest round. In the 1990s, elections in India have witnessed a high level of churning of votes. While the overall vote shares of the major parties do not undergo any dramatic change, a very large number of voters change their voting preferences from one election to another. In this respect, however, Elections '99 have proved very different from the previous two elections. All the major parties have been able to retain the support of a large chunk of those who voted for them in 1998.
But one must not read too much into these figures. These are provisional 'recall'-based figures that are liable to being scaled down after checking out carefully. Also, the recent round of elections were held within 20 months of the previous elections and without any major change in the pattern of electoral alignments at the all-India level. Yet the figures do indicate a greater crystallisation of voter preferences. It is not clear if this phenomenon will endure and lead to a deeper institutionalisation of the party system in the face of the social churning that is taking place. But it does suggest that the process of creation of a social bloc may have reached a certain stage of completion.
A closer look at how votes changed hands between 1998 and 1999 does not support the suggestion made by some commentators that voters are drifting towards a bipolar choice. Most of the gains made by the BJP-led front came from the erstwhile United Front and the allies of the Congress(I). But these votes did not flow directly to the BJP: they accrued to its allies. If anything, in a direct exchange of votes, the Congress(I) gained from the BJP more than it lost to it (11 percentage points gained, 6 percentage points lost). The Bahujan Samaj Party and the Left parties retained the support of a very high percentage of their voters, but largely because their base is not expanding to a considerable extent. The U.F's votes were fragmented and went either to the BJP's allies, the Congress(I) or to several regional parties. This fragmentation prepared the ground for the creation of the new social bloc by the BJP.
A look at the division of votes along caste and community lines gives one an idea of the shape of this new social bloc. The ten-fold classification used here is not the most refined one, and certainly does not do justice to State-level configurations, but it does give a more elaborate description of the community basis of voting than is usually available.
The first point to note is the difference between the support base of the BJP and that of its allies. Together, the BJP and its allies secure the support of 60 per cent of upper-caste Hindus and 52 per cent of the dominant Hindu peasant castes (which are not classified as Other Backward Classes) such as Jats, Marathas, Patidars, Reddys and Kammas. But between the BJP and its allies, the relative share of the allies goes up as one moves down the social hierarchy from the upper castes to the lower OBCs. This is a new pattern, as compared to 1998 and earlier elections. Since the 1998 elections, the BJP-led alliance's share of votes from among the second, third and the fourth categories has gone up by nearly 10 percentage points each. And the BJP's own share among these "middle castes", as compared to those of its allies, has declined from about two-thirds to one-half or even less. The lower the category, the lower the BJP's contribution to the vote share of the alliance. It is not that voters have shifted from the BJP to its allies in the 1999 elections. It is just that regional and political expansion has brought in new middle-caste parties such as the DMK, the Janata Dal (United) and the Telugu Desam Party and clarified the tendency inherent in the BJP's rise to power.
The vote share of the BJP-led alliance drops among Adivasis and Dalits. Although both these figures are higher this time as compared to the 1998 elections, it is clear that these groups are not the primary constituents of the new social bloc. The figures go down even further when one turns to the minorities. The BJP has maintained a fair share of the votes of Sikhs ever since the community became disillusioned with the Congress(I) following the anti-Sikh riots in 1984; but the BJP inspires little confidence among Muslims and Christians.
One needs to look at subtle changes here. Throughout this decade, the vote share of the BJP-led alliance among Muslims has grown, even if only slowly; in the 1999 elections, the share reached a two-digit figure. At the same time, the BJP's own share of votes of Muslims has declined compared to 1998, and is concentrated among Ashrafs. The increase in the alliance's vote share among Muslims is entirely owing to allies such as the TDP, the Trinamul Congress and the DMK. Similarly, in the BJP-led front, the allies account for more than half of the votes of Christians.
The Congress(I), on the other hand, improves its vote share as one moves down the social hierarchy. It has the support of only 21 per cent of upper-caste Hindus, but its vote share among the dominant peasant castes is about 31 per cent. Its vote share declines among the OBCs, but its allies' support base among these sections makes up for it. Compared to 1998, the Congress(I) has lost support among the dominant peasant groups such as Jats and Marathas, but has gained greater support among both the categories of OBCs. The Congress(I) gets the biggest share of the votes of Adivasis, Dalits, Muslims and Christians. Compared to 1998, the Congress(I)'s vote share among Adivasis has declined by four percentage points; there appears to have been a decline in its vote share among Dalits as well, but its allies make up for this.
Among the two sections of Muslims (upper biradari and lower biradari), the vote share of the Congress(I) and its allies has gone up by about 10 percentage points; the Congress(I)'s gains are principally among the upper biradari Muslims, while much of the gains from among lower biradari Muslims came from allies, mainly the Rashtriya Janata Dal (RJD). This indicates a general return of the minorities' faith in the Congress(I); this is confirmed by the fact that the party's vote share among Christians and Sikhs has increased by about eight percentage points. A small but significant section of the Sikh community has returned to the Congress(I) after a long time.
However, the Congress(I) has not been able to consolidate the support of the lower sections (except in the case of Muslims and Christians) the way the BJP and its allies have consolidated the support of the upper castes. The Left parties and the BSP challenge the Dalit vote of the Congress(I), while a host of regional parties such as the Samajwadi Party, the Nationalist Congress Party and the Tamil Maanila Congress challenge the OBC and Muslim vote of the Congress(I).
While religion is a marker of social community, especially for the minorities, political loyalties in India are not defined by religious practice. This stands in sharp contrast to several European democracies where church-attendance or otherwise explains political and ideological cleavages. A classification of Hindus in these surveys by their religious practice or otherwise does not give very firm clues to their voting pattern. The vote for both the BJP and the Congress(I) shows a mild and positive relationship to the regularity of worshipping, while it is the other way round for the Left parties and the BSP. But then that is also because there is a relationship between belonging to the upper and middle castes and worshipping regularly. Interestingly, 41 per cent of the Hindus who vote the Left worship regularly. The same figure is substantially lower for the BSP.
From caste-based and community-based cleavages, we can now turn to the economic dimension of the emerging social bloc, its class structure. The table shows how clearly the BJP's vote is related to class. The poorer the voter, the lesser are the chances of voting for the BJP. The BJP vote share among the poorest of the population is about one-third of what it gets among the upper classes. The BJP's allies do not share this profile, for they tend to do better among the middle category. In that respect there is no significant difference in the BJP's support base as compared to 1998.
If the Congress(I) and its allies are taken together, it appears that their profile is the opposite of that of the BJP. But when the allies are seen separately, it is clear that two of them, the All India Anna Dravida Munnetra Kazhagam (AIADMK) and the RJD, have a pronounced lower class support base, while the Congress(I) support to them does not make for any clear pattern. This absence of any pattern is owing to the fact that the class patterns of Congress(I) voting in different States go in different directions and eventually cancel each other out at the national level. As is to be expected, the Left parties and the BSP enjoy greater support among the poorest voters.
Till now we have seen the caste effect and the class effect independent of each other. But as every student of Indian society knows, there is a considerable convergence of the two systems of stratification at the top and the bottom. In other words, the upper caste people tend to be mainly upper class and people belonging to the lowest caste in the Hindu social order tend to be among the poorest. Yet it cannot be said that there is complete overlap here. Class and caste cleavages represent two independent yet overlapping principal cleavages of the Indian social order. We have seen that the BJP tends to attract the voters from the upper segment of both these divisions.
The data also shows that the caste hierarchy and the class hierarchy reinforce each other in contributing to support for the BJP. Among the upper castes or the dominant peasant castes, for example, the vote share for the BJP goes down as we move from the upper classes to the lower classes. The same is true for different classes among Adivasis and Dalits. The pattern for the OBCs is not very neat at the lower ends but supports the basic reading offered here. Conversely, if we focus on the lower classes and move from Dalits to upper caste, the vote share for the BJP tends to go up. Accordingly, the highest chances of voting for the BJP obtain in the group that is both upper class and upper caste, and the lowest among the lowest class Dalits.
This convergence of class and caste in voting for the BJP can be understood more clearly by drawing a diagonal line across the caste-class divide. The line separating those who vote BJP in greater numbers than average and those who do not is also roughly the line separating the socially and economically privileged from the underprivileged. Those included in the "upper strata" of this caste-class divide include all the upper caste Hindus irrespective of their class. As we come down the social ladder, we exclude the lowest class and then the lower classes as well from the "upper strata" thus defined. Among the Adivasis, only those who belong to the two upper classes qualify for the "upper" caste-class strata, and among Dalits, only the highest quality thus. This line excludes all the minorities irrespective of their class. If we divide the entire electorate along these lines of "upper" and "lower" caste-class strata, we can see a composite picture of the new social bloc.
The table presents the proportion of each party's voters from the two strata, rather than the vote share for different parties as in all other tables. In all, 45 per cent of our electors fall in the "upper strata" and 55 per cent in the "lower strata". The BJP, however, draws as much as 69 per cent of its votes from these 45 per cent voters. Its allies are a little less lop-sided, but they also draw more than a proportionate share of votes from this privileged group. All the remaining parties depend more on the lower strata for their votes. This is, of course, more true of the Left parties and the BSP than of the Congress(I) .
This social bloc formed by the two principal cleavages of caste and class is reinforced by several auxiliary divisions that more or less overlap with the caste-class divisions. A look at the educational background of the voters shows that the education effect works in the same direction and to a similar degree as class. As one goes up the educational ladder, the odds of voting for the BJP go up very sharply. The votes for the BJP's allies are fairly evenly distributed while there is a clustering in the middle for the Left parties. This pattern had already been established before the last elections and has remained more or less unchanged. While a good deal of the education effect is accounted for by the social class factor, for education is largely a function of the class someone is born in, there seems to be a milder and independent education effect after controlling for caste-class. In both the upper and the lower strata, the Congress(I) vote falls and the BJP vote rises with increase in levels of education.
The effect of reading newspapers, listening to the radio and watching television is very closely related both to education and to class. No wonder, the BJP does better among those who are more exposed to the media. In this case the BJP's allies too seem to be gaining significantly from media exposure. As in the case of education, the media exposure effect is not entirely explained by caste-class background. Within the upper stratum, differences in media exposure produce significant differences in voting for the BJP and its allies. Clearly, greater education and greater media exposure by themselves seem to be drawing voters away from the Congress(I). The same is not true of the Left parties, the BSP and other regional parties, who also get more than the average level of vote among the under-privileged.
Right from its inception, first the Jan Sangh and then the BJP has drawn greater support from the urban areas. The same pattern continues, although the gap has kept changing from four percentage points in 1991 to 12 points in 1996, before plateauing to six points in 1998 and seven this year. This gap does not indicate, however, that the BJP has no rural base. On the contrary, of every hundred votes it gets, as many as 73 come from the villages, only slightly below what the Congress(I) draws from the villages. Earlier, the Congress(I) used to get a higher level of votes in rural areas. In this decade, however, Congress(I) has also done marginally better in rural areas. The main political base of the Left parties is in the rural areas. But over the years, the gap between their rural and urban votes has increased because of their shrinking urban appeal in West Bengal. Urbanity by itself does not seem to introduce a new dimension to political preference. Within the upper and lower strata, the vote for BJP does not vary much.
Unlike in Western democracies, age or generation identity has not been a significant factor in Indian elections. The sudden rise of the BJP in the early 1990s had opened for the first time a noticeable generational divide. Among the youngest of voters, the BJP secured four percentage points more votes than among the oldest. The Congress(I) showed a contrary pattern. By 1996 the gap had narrowed to 2.5 percentage points and this time it virtually disappeared. It seems the gap was more a result of the BJP's newness in many parts of the country than any generational divide. The BJP does marginally better among the youngest and the Congress(I) among the middle aged. The Left has consistently done a little worse among young voters as compared to the middle aged.
Traditionally, the Congress(I) used to witness a fairly big gender gap, with support for it among women being five to six percentage points more than among men. The sudden decline of the Congress(I) and the rise of the BJP had reduced that gap considerably by the beginning of this decade. The gap started increasing again in 1996 when the Congress(I) enjoyed a 1.2 percentage point lead. By last year the gap had doubled to 2.7 points in favour of the Congress(I) and 4.7 points against the BJP. While in the case of the BJP the gap has decreased a little, the gender gap for the Congress(I) and its allies stands at five percentage points, as high as during the good old days of Congress(I) dominance. In the case of the Communist Party of India (Marxist) and the BSP, the gender profile has not undergone any change all these years.
It is not clear how the revival of the gender gap should be interpreted. It is all too easy to read it as conclusive evidence of a revival of the Congress(I) or to dismiss it as yet another sign that the Congress(I) gets the votes of the uninformed and the less politicised. At the same time, it may be too premature to link it to women's attitudes to the policies of the Congress(I) or the BJP. It calls for more careful research, especially because this is happening at a time when women's political participation is growing. At this stage we can only view it as another independent dimension of social hierarchy, where the privileged section supports the BJP.
To sum up, the 1999 electoral verdict points to a fundamental reworking of the social basis of party politics in India. The first four decades of the Congress(I) system in one form or the other were characterised by political mobilisation that cut across fundamental social cleavages. In its various incarnations and local variants, the rainbow social coalition of the Congress(I) did not include all the sections quite equally. But it did not follow any one faultline of society. The decline of the Congress(I) and the rise of the BJP to power created the possibility of a new kind of cleavage-based politics, one that draws on the overlap of cleavages based on caste and class. It is not caste-based or class-based in any simple sense: it is woven around the ideology of nationalism and involves a reworking of Jati and sectional divisions. Nonetheless, it is built around a "master cleavage" of caste-class privileges. Its end-product is a new social bloc, one with soft edges and blurred boundaries. Verdict '99 marks the arrival of this new social bloc.
To be sure, the arrival does not guarantee its durability or even survival, let alone its continued electoral success. The new social bloc is still quite fragile and yet to jell. In many ways it is a very artificial product, more a result of a highly skillful working of the logic of our electoral system and caste-community configurations in their regional setting by the BJP leadership than a harmonious coming together of various groups. As noted above, there is a definite tension among the social profiles of the BJP and its allies. But at present it helps the BJP bring groups to its new social bloc that would not have otherwise come its way. It is not clear how well the BJP will be able to manage this tension. It is also not clear if the party will succeed in mobilising enough numbers from outside this group so as to ensure the creation of an electoral majority out of a numerical minority.
This analysis naturally begs the question: if the privileged can form a bloc, why not the underprivileged? Do the recent elections and political indicators point in that direction? In a sense the decline of the "catch-all" Congress umbrella opened up the possibility of the political consolidation of the underprivileged as it did for the privileged. This decade has enriched that possibility, for we have witnessed a participatory upsurge of the lower orders of society. One needs to take note of the extraordinary fact that India is perhaps the only contemporary democracy where the poor and the socially underprivileged participate in politics more than the elite.
At the same time, various factors have worked against the realisation of this potential and resulted in the fragmentation of what could have become a counter-bloc. First, it is always more difficult to consolidate the underprivileged, for their access to information and action is limited. Second, the regionalised nature of our polity works against the coming together of these sections on a national political platform, leaving it open for the BJP to incorporate a group like the Pattali Makkal Katchi through localised negotiation. Finally, the absence of an organised party-political nucleus that would initiate such a consolidation prevents the potential from turning into a reality.
Can the Congress(I) perform that role? On the face of it, the findings presented above might suggest that the Congress(I) is already well on its way to creating a counter-bloc. It is after all the party that gets the largest support from among the underprivileged. At the same time we need to remember that the Congress(I) does not draw the lower strata of caste-class as fully as the BJP does for the upper class-caste. The national level figures for the Congress(I) hide various contrary trends at the State level. The Congress(I) draws the underprivileged towards it in Karnataka, has an unclear profile in Andhra Pradesh and represents the better-off in Kerala. Besides, when the Congress(I) gets the votes of the lower strata it comes more by default than by design.
There is a residual quality to the Congress(I) vote. The social support it gets is more often than not merely the mirror image of the social profile of its opponent. The creation of a counter-bloc of the underprivileged would need more than a happy coincidence. It would require a painstaking building of social alliances and political coalitions, within or without the party. It would also require forging a new vision and overhauling the organisation to allow it to act as the vehicle of this historical process. For the existing Congress(I) party, it is a very tall order indeed.
COMMents
SHARE