Choosing thy neighbour

The very process of development and change in India may be generating new forms of social and economic competition that manifest themselves in terms of social bias

Popular debate around social biases in India is structured around two competing narratives. One view holds that as an urbanising country with rapid economic growth over the past few decades, the importance of ascriptive identities such as caste and religion is gradually eroding. An opposing view holds that these biases have remained resilient in India, even in the face of substantial economic development and increasingly heterogeneous cities.

Yet, such a simple dichotomy understates the complexity in characterising social biases in India. New forms of bias may emerge while other forms fade away. While social biases often result from prejudice or chauvinism, they may also result from legitimate apprehensions about, or threats from, another social group. In order to develop a deeper understanding of the profile of social biases in India, we analyse new data from the Lok Surveys, taking advantage of both the scale and the geographic spread of the sample. Before describing our results, we note that any survey-based analysis of social bias is necessarily fraught with difficulties — questions about bias are sensitive and respondents are often unwilling to admit to their biases. Furthermore, there is no universally accepted tool used to measure bias.

Identity of neighbours

Rather than relying on complex typologies that can be impacted by preconceived notions, we focus our analysis on a simple topic, which we believe represents a core form of social bias: differences in preferences for the identity of one’s neighbours. These preferences capture important dimensions of social structure. They involve beliefs about how different social groups affect social solidarity in a neighbourhood, as well as apprehensions about interacting with different social groups. To uncover social biases in preferences for neighbours, each of our respondents was asked the following question: Would you be against having a family of (another identity group) as a neighbour?

“It is the middle class group that accounts for much of the social bias in preferences for the identity of one’s neighbours”

The part of the question in brackets was replaced by a randomly generated prompt. First, we randomised whether the respondent would receive a prompt for religious or caste bias. Then, based on this first stage, we asked about a social group different from that of the respondent. For instance, a Hindu respondent slotted to receive a religious bias prompt might be asked about a Muslim neighbour, and a Muslim respondent might be asked about a Hindu neighbour. Similarly, an upper caste individual slotted to receive a caste bias prompt could be asked about Other Backward Classes (OBCs), Scheduled Castes (SCs), or Scheduled Tribes (STs). In this manner, we generated data on a complex set of social interactions in Indian society. It should be noted, however, that the results are based on correlations, not multivariate analyses.

Overall, 27 per cent of the sample population directly admitted that they were against having a neighbour from a different religious or caste community. While this number seems high, we cannot ascertain whether this is part of an increasing or decreasing trend in social biases over time because, to our knowledge, this question has never been asked before in a large nationally representative survey in India.

There is, as to be expected, significant variation in levels of social bias across States. Punjab displays the highest level of social bias, with 36 per cent of respondents displaying an aversion to living near those of another caste or religion, while the post-split Andhra Pradesh (Seemandhra) displays the lowest levels of social bias, with 12 per cent of respondents displaying aversion. Tamil Nadu, Karnataka, Chhattisgarh and Rajasthan also had relatively high rates of social bias (each with over 30 per cent of respondents indicating an aversion).

Social bias permeates all segments of Indian society. The education level or wealth of respondents had little impact on whether they would report social bias. Furthermore, there was little difference between how village dwellers and city dwellers responded to the question; 28 per cent and 27 per cent of rural and urban respondents, respectively, indicated social bias. These findings suggest that urbanisation and improved access to education may not reduce social bias.

The majority of the variation in social biases is seen between specific identity groups. SC and ST populations demonstrated a greater aversion to living near upper castes than to living near other marginalised communities, including Muslims. In total, 29 per cent of SCs indicated a social bias against upper castes, as compared to 24 per cent towards STs, and 38 per cent of STs indicated a social bias against upper castes, as compared to 24 per cent against SCs. Given the reality of caste hierarchies, perhaps marginalised communities are apprehensive that traditionally dominant communities will discriminate against them or hurt social solidarity in their neighbourhoods. A similar story may explain why Muslims display somewhat greater aversion towards Hindu neighbours (31 per cent) than Hindus do towards Muslim neighbours (27 per cent), especially considering that much of that gap is due to the relatively high rate of “lower caste” Muslims who were against living near a Hindu family (8 per cent higher than “upper caste” Muslims).

Not all social biases are driven by marginalisation. Though our survey cannot gauge the intensity of these preferences, upper caste respondents were more likely to say they did not want to live near OBCs than any other group. Overall, 34 per cent of upper caste Hindus admitted preferences against OBC neighbours, as compared to 26 per cent against SCs and 23 per cent against STs. A politically ascendant OBC population has begun to challenge high caste dominance in many social spheres, creating greater competition for resources. Based on this data, we conjecture that social bias may also be generated from threats to power and intensifying economic and social competition.

In order to understand the role of social and economic ascendance on preferences for neighbours, we also investigated the relationship between middle class identification and social bias. In the previous piece in this series (“Being middle class in India,” December 9, 2014), our colleagues Devesh Kapur and Milan Vaishnav demonstrated that 49 per cent of all Indians, including people from all segments of Indian society, self-identified as “middle class.” They were more optimistic about the economic conditions of their family and the nation as a whole as compared to those who did not identify as middle class. The self-identified middle class was also more likely to report that their family had experienced social mobility within the last generation.

However, it is precisely this socially mobile group that accounts for much of the social bias we observe in our sample. Among those who do not identify as middle class, only 17 per cent of respondents said they did not want a neighbour from a different community. However, among those who perceive themselves as middle class, 39 per cent indicated social bias against a religious or caste community. The reasons for these large differences are not immediately clear. It is likely, however, that there is something fundamental about the construction of middle class identity that lends itself to social bias, as described in the earlier piece.

Causes for bias

Based on our data, we suggest two very different causes for bias in one’s preferences for neighbours. Marginalised communities display higher levels of social bias against traditionally dominant communities, perhaps as a reaction to historical stigmatisation and concerns for social solidarity. But traditional social marginalisation is not the whole story. Upper caste Hindus now report the most social bias against OBCs, instead of groups lower on the caste hierarchy. Those identifying as the Indian middle class display much higher levels of social bias than those who do not identify as such. We surmise that the very process of development and change in India may be generating new forms of social and economic competition that manifest themselves in terms of social bias. A modernising India may trigger the erosion of certain traditional hierarchies while, at the same time, opening the way for new cleavages based around social and economic contestation.