Why Voters in India Vote the Way They Do

Understanding Indian voter behaviour requires abandoning simple explanatory frameworks. Indian elections — particularly Lok Sabha elections — are not single-issue contests nor do they unfold uniformly across 543 constituencies. A voter in a Varanasi constituency is making a calculation about caste, religion, development, and Modi's personal leadership that is structurally different from a voter in a Chennai constituency where Dravidian cultural identity, state welfare delivery, and anti-BJP sentiment are the dominant axes. At the same time, certain patterns run across India's electoral geography consistently enough to characterise Indian voting behaviour at the aggregate level.

Representational Image: Why Voters in India Vote the Way They Do

Academic research, national election studies by CSDS/Lokniti, and post-election analysis from Carnegie Endowment converge on a picture in which Indian voters are neither purely rational economic actors voting on income and development outcomes nor purely identity-driven bloc voters following caste or religious leadership. They are political actors who weigh multiple considerations simultaneously — some of which (caste identity, religious identity) are relatively stable, and others of which (development performance, anti-incumbency, local candidate quality, specific welfare benefits received) are variable. Understanding which considerations dominate in a given election cycle, in a given state and constituency type, is the core challenge of Indian electoral analysis.

What You Need to Know

• CSDS/Lokniti national election studies consistently show that "development" and "employment" top voter priority lists in survey data, but these preferences do not deterministically predict outcomes — voters may cite development as important but vote on identity lines in practice; the gap between stated preference and revealed preference is a feature of voter survey research globally.
• Anti-incumbency is a powerful and well-documented driver in Indian state elections: incumbents face higher hurdles in re-election than challengers across most states; the "anti-incumbency factor" represents voter frustration with governance failures, unfulfilled promises, and bureaucratic corruption that accumulates during terms; Bihar under Nitish Kumar reversed this with sustained development delivery.
• The "welfare voter" is a documented phenomenon — beneficiaries of PM-KISAN, PM Awas, UJJWALA, and Ayushman Bharat show higher BJP vote shares than non-beneficiaries in most analyses; Carnegie Endowment confirmed the "welfare-nationalism" combination is BJP's most durable electoral formula.
• Local candidate quality matters: survey data consistently shows that a significant minority of Indian voters (15–25% in most surveys) report that the local candidate's personal reputation, accessibility, and perceived competence is a significant factor in their vote; this creates variance around national and state-level swings in individual constituencies.
• First-time voters (18–22 year olds) have consistently shown higher BJP support than older cohorts in recent elections, according to CSDS data; this demographic skew reflects the generation that grew up during the Modi era, with BJP as the dominant party of their formative political experience.

How It Works in Practice

1. The economic voting model and its limits: Rational choice theory predicts that voters reward governments for economic growth and punish them for economic decline. India's 2024 election partially confirmed this — unemployment concerns (India's youth unemployment rate at 23.2% according to a 2022 World Bank report cited by Wikipedia) contributed to BJP underperformance despite GDP growth; but the correlation is imperfect because identity factors, welfare receipt, and political narratives complicate pure economic voting.

2. Identity voting and its rationality: Caste and religious identity voting is often described as "irrational" by development economists who want voters to focus on policy performance. But identity voting reflects rational expectation: a candidate from your caste community is more likely to respond to your grievances, allocate resources to your community, and represent your interests in state and national institutions. This rationality means identity voting is not simply a "distortion" to be corrected by education — it reflects real institutional dynamics of patron-client governance.

3. The "Modi factor" in national elections: 2014 and 2019 elections had a significant "presidential" quality — voters were choosing Modi as Prime Minister more than BJP as a party. Lokniti-CSDS data showed 47% preferred Modi as PM in 2019 vs 14% for the Congress alternative. The 2024 decline to 41% (with Congress leader rising to 27%) reflects the first partial erosion of the personal vote premium; it contributed to BJP losing 63 seats despite similar national vote share.

4. State-national divergence: Indian voters routinely vote differently in state and national elections, even held simultaneously. The 2024 data showed Congress winning Karnataka state (May 2023) while BJP won most of Karnataka's Lok Sabha seats (June 2024); Telangana state went to Congress (December 2023) while Congress did well in Lok Sabha too; but Bihar state stayed with NDA (November 2025) consistently. Voters assess state and national governments differently, using different criteria.

5. The "underdog effect" and sympathy waves: Indian electoral history shows that underdog narratives can mobilise voters in specific elections — Indira Gandhi's "Garibi Hatao" (1971), Rajiv Gandhi's sympathy wave (1984), Modi's "change" narrative (2014), and the 2024 opposition's "Constitution in danger" narrative each represent successfully mobilised emotional-political frameworks. Conversely, overconfidence can suppress supporter turnout as happened to BJP in 2024 when its "400 paar" campaign may have reduced mobilisation by signalling inevitable victory.

What People Often Misunderstand

• Voters are not uninformed or irrational: India has one of the world's highest voter turnout rates (66% in 2024) among eligible voters; Indian voters make complex multi-factor decisions; the finding that poorer voters vote at higher rates than richer voters in India is opposite to the pattern in most Western democracies and reflects genuine political engagement.
• "Vote buying" does not deterministically produce outcomes: Cash distributed before elections does not simply "buy" votes; voters take the cash and vote according to their preferences; the correlation between pre-election cash distribution and vote outcomes is positive but weak, as documented in academic studies of Indian vote-buying.
• Anti-incumbency is state-specific: Some states show strong anti-incumbency (Rajasthan, which alternated Congress-BJP for 30 years until BJP's 2023 win); others show weak anti-incumbency (Tamil Nadu alternated DMK-AIADMK for decades, showing incumbent pattern rather than anti-incumbency).
• Urban-rural divides in voting behaviour are real but not absolute: Urban voters tend to show slightly higher BJP support (in national elections) and slightly more issue-based voting; rural voters show stronger community network effects and welfare scheme response; but both urban and rural voters use multiple criteria, and the divide is a tendency rather than a binary.
• Polling data in India is less reliable than in Western democracies: India's polling industry has a poor track record of predicting election outcomes accurately; the 2019 and 2024 elections both produced results significantly different from pre-election polls; this is partly methodological (sampling rural India at scale is difficult) and partly because last-minute shifts and election-day dynamics are hard to capture.

What Changes Over Time

The rise of digital information — particularly WhatsApp-mediated political messaging — has accelerated the pace at which political narratives shift voter perceptions between election announcement and polling day. The 2024 "Constitution in danger" narrative — mobilised by opposition parties and resonating specifically with Dalit voters in northern India — shifted enough votes in UP to cost BJP 29 seats; this narrative was partly WhatsApp-driven and was not captured in pre-election polls. As digital penetration deepens and AI-generated content proliferates, the speed and scale of narrative shifts during election campaigns will continue to increase — making the gap between polling date and result more significant.

Sources and Further Reading

• Carnegie Endowment — The Resilience of India's Fourth Party System: https://carnegieendowment.org/research/2024/09/india-election-bjp-party-politics
• SSRN — Caste and Constitution: Dalit Voting Patterns 2024: https://papers.ssrn.com/sol3/Delivery.cfm/5361057.pdf?abstractid=5361057&mirid=1
• Al Jazeera — How will votes be counted: https://www.aljazeera.com/news/2024/6/2/india-lok-sabha-election-2024-how-will-votes-be-counted
• CASI — Eswaran Sridharan on India's Coalition Politics: https://casi.sas.upenn.edu/iit/rohan-venkat-eswaran-sridharan-interview-2024

(This series is part of a long-term editorial project to explain the institutions, ideas, actors, and power structures that shape political life in India. Designed as a 25-article briefing cluster on Indian Politics, Elections & Political Power, this vertical examines how electoral democracy functions in practice — from voting systems, political parties, caste coalitions, campaign finance, and the Election Commission to ideological movements, opposition politics, coalition-building, and the exercise of political power at both national and state levels. Written in an accessible format for diplomats, investors, researchers, academics, journalists, students, policymakers, civil society organisations, and international observers, the series seeks to explain not only how India's political system is formally structured, but also how political competition, representation, and governance operate in reality. This is Vertical 5 of a larger 20-vertical knowledge architecture being developed by IndianRepublic.in under the editorial direction of Saket Suman. All articles are protected under applicable copyright laws. All Rights Reserved.)