Political Parties and Voting Behaviour

The sociologically distinctive fact about India is the strength of state-level parties with no national presence, which follows from the regional plurality described in [The Pluralistic Nature of Indian Society].

Voting behaviour

Why it is a sociological subject

A vote is cast alone and in secret. If choice were purely individual, votes would scatter and there would be no pattern to study. In fact they cluster by social characteristic, in every country studied, and the clustering is stable enough to be predicted within limits. That regularity is a social fact in Durkheim's sense, and it is why voting is in a sociology syllabus at all.

The determinants studied in India

The correct treatment names the factors and stops there, without asserting how any group votes.

Caste. The most studied factor in Indian electoral sociology. It works in several distinct ways which should be distinguished: as a basis on which candidates are selected for a constituency; as a network through which campaigning is conducted; as a source of identification with a candidate; and, where a caste group is numerically significant and organised, as a bloc whose leaders can negotiate. Rajni Kothari's formulation is the standard reference point: the interesting process is not only the caste basis of politics but the politicisation of caste, in which caste identities are reshaped by their use in electoral competition.

Religion. Both as identification and as an issue.

Region and language. Which is why state parties exist.

Class and economic position. Occupation, landholding, income and access to public benefits.

Rural or urban residence, which affects both turnout and the issues that matter.

Age, with a very large young electorate in India.

Gender. The most striking recent finding, and one that can be stated confidently because the Election Commission publishes turnout by sex: the historic gap between male and female turnout in Indian general elections has narrowed very substantially, and in recent elections female turnout has been close to and at times above male turnout. That is a well-documented change and a safe thing to write.

Education and information.

Party identification and leadership, including the personal appeal of a leader, which can cut across every social factor above.

Local candidate factors: the candidate's own standing, accessibility and record.

Incumbency, and the well-attested Indian pattern of anti-incumbency, which is a tendency rather than a law.

The theoretical models

Three models are conventionally given and each explains part of the picture.

The sociological model. Vote follows social location: the groups a person belongs to shape their choice through shared interests and social pressure. Its strength is the observed clustering; its weakness is that it cannot explain change, since social characteristics move slowly and votes move fast.

The psychological or party-identification model. Voters develop a durable attachment to a party, acquired through socialisation in the family, which then filters how they see issues and candidates. Its strength is explaining stability; its weakness is that party identification is weaker in India than in the countries where the model was built.

The rational choice or economic model. Voters assess parties by expected benefit and vote for the one that offers most. Its strength is explaining swings and the salience of welfare delivery; its weakness is that the cost of acquiring information is high and the chance of one vote being decisive is negligible, which makes the calculation questionable.

The correct answer says that no single model explains Indian voting, and that the factors interact: caste supplies the network, the leader supplies the appeal, and the assessment of delivery supplies the swing.

A worked example

A rural constituency in Maharashtra with roughly two lakh voters.

A sociologist studying it would ask: what is the caste composition, and are any groups numerically dominant; what is the pattern of landholding, since that decides economic dependence; which cooperative institutions exist and who controls them, since in Maharashtra sugar cooperatives and credit societies have been political structures as much as economic ones; how many voters are recent migrants; what is the turnout by sex; and who the local party workers are and what services they provide between elections.

Notice what the list does not contain: any assumption about how a caste will vote. The sociologist's questions are about structure and mechanism, not about predicting outcomes, and this is the difference between electoral sociology and psephology.

The criticisms

1. It is difficult to measure. The ballot is secret, so the data comes from surveys with all their problems of sampling, recall and the tendency of respondents to report what they think is acceptable, which is discussed in [The Interview].
2. Ecological fallacy. Inferring how individuals voted from the results of an area where they live is a classic error: a constituency with many voters of a certain group and a large vote for a certain party does not establish that those voters cast that vote.
3. Determinism. Treating a social characteristic as deciding a vote denies the voter's judgment, and the evidence of large swings between elections with an unchanged social composition contradicts it.
4. Categories are unstable. Caste and community categories used in analysis are themselves shaped by politics, which is Kothari's point, so the explanatory variable is partly an effect of the thing it explains.
5. The list of factors is not an explanation. Naming caste, class and region does not say how they combine, and much writing stops at the list.

Quick revision

1. A party seeks to capture power; a pressure group seeks to influence it. That is the defining difference.
2. Functions of parties: interest aggregation, recruitment of leadership, political socialisation, opinion formation, providing government and opposition, and linking citizen to state through local brokerage.
3. Types: by ideology; national and state parties, a distinction recognised by the Election Commission; cadre and mass; broad-appeal and group-appeal.
4. Voting is studied sociologically because votes cluster by social characteristic, which is a social fact.
5. Determinants: caste, religion, region and language, class, rural or urban, age, gender, education, party identification and leadership, candidate factors, incumbency.
6. Kothari: not only the caste basis of politics but the politicisation of caste.
7. Gender: the male and female turnout gap in Indian general elections has narrowed very substantially, and female turnout has recently been close to or above male turnout.
8. Three models: sociological, party identification, rational choice. None alone explains Indian voting.
9. Criticisms: secrecy makes measurement hard, the ecological fallacy, determinism, unstable categories, and a list of factors is not an explanation.

Test yourself

1. Distinguish a political party from a pressure group. A political party seeks to capture governmental power itself, principally by contesting elections, and takes responsibility for governing across the whole range of public questions. A pressure group does not seek office; it seeks to influence whoever holds it, usually in relation to a narrow set of interests, and it can therefore work with any government.

2. State four functions of a political party. Interest aggregation, bundling the innumerable demands of a society into a few packages that can be voted on; recruitment of political leadership, since parties select who will hold office; political socialisation and the formation of public opinion; and linking the citizen to the state, which in India frequently takes the concrete form of local workers assisting people with documents, entitlements and access to officials.

3. Why is voting behaviour a sociological subject? Because although a vote is cast individually and in secret, votes cluster systematically by social characteristic: caste, religion, region, class, age, gender, education and residence. That regularity persists although the individuals change, so it cannot be explained by facts about any particular voter. It is a social fact in Durkheim's sense and therefore a proper object of sociological study.

4. What did Kothari mean by the politicisation of caste? That the relationship between caste and politics runs both ways. Caste supplies politics with networks, identities and candidates, but electoral competition also reshapes caste: it encourages the formation of wider caste federations, produces associations and leaders where none existed, and gives old identities new political content. The categories used to explain voting are therefore partly a product of the electoral process they are used to explain.

5. What is the ecological fallacy in the study of voting? The error of inferring the behaviour of individuals from data about the area they live in. That a constituency contains many voters of a particular community and returned a large vote for a particular party does not establish that those voters cast that vote; the correlation is at the level of the area, and the individuals within it may have voted in any pattern consistent with the total.