Voter Turnout

Spatial Patterns and Voter Turnout

By: Sulagna Chattopadhyay
The strongest feature of Indian democracy is the high rate of participation of citizens in elections. Voter turnout, political awareness and participation of women are in elections constantly changing and are affected by various factors like literacy, economic development and influence of the media.
Magazine Articles Policy

India conducts fair and democratic elections, where voters exercise free will through constitutionally mandated universal adult suffrage.  Voting behaviour and its Voter turnout is closely linked with the attitude of the populace towards voting. It is therefore imperative to understand how the masses respond to voting, which could serve as an impactful indicator to how democracy may evolve in the future (Banerjee, 2014). There is, however, a larger question that looms – does voting really involve free will?

Free will does not exist. Even though the voter feels that he has not been coerced into voting for any single party, a gamut of considerations have actually worked their way into his decision. Poverty, level of education, employment, gender, exposure to media and general awareness about the political process all play a significant role in the voting behaviour and voter turnout.

Spatial dimensions of Voter Turnout

Voter turnout in India has been consistently high for the last seven decades, touching 67 per cent in the 2014 general elections (ECI, 2014). India’s democracy is robust with a steady, and in some states, rising voter turnout, in sharp contrast to the democratic regimes of developed nations (Taylor, 2007). A case in point is the USA, where voter turnout has dropped to 56.5 per cent in 2016 as compared to 58.6 per cent in 2012 (National Journal Leadership council, 2016).  The elections in India are the largest humanly organised event in the world and as per the 2014 General Elections data (ECI, 2014), involving over 554 million voters, 15,415 million election officials, 1.7 million EVM machines (Ford, 2014), more than 687,402 polling booths and 543 constituencies. Despite this, the process is run with a smoothness that is truly appreciable, giving elections a distinctive character in the sub-continent.  With the introduction of the EVM machines in 1982 (ibid), the use of the party symbol corresponding to each party name, placing the EVM machines away from the gaze of officer or bystander and marking of the fingers with indelible ink to avoid voter fraud (Banerjee, 2001) have made the process credible for the millions that appear unfailingly each time at the polling booth.

The voter turnout presents an interesting spatial array (Fig. 1). The east and north-east clearly takes it all. These states have been presenting a consistently high voter turnout. Theorists claim that voter turnout is usually higher if the people are looking for a change. Nagaland, Tripura, Sikkim, West Bengal, Assam and Manipur have shown voter turnouts that hover around a humongous 80 per cent, followed by the southern states where voter turnout hits a robust 65-75 per cent bracket. Perhaps anti-incumbency may not be the only factor that drives a high voter turnout. In the case of north and central Indian states, voter turnout, despite spiking in the 2014 general elections, has generally shown a near 50 per cent trend. A few states however deserve mention, especially in the context of the 2014 data, where a historic high in voter turnout has been achieved—Goa, followed by Gujarat and Rajasthan (ibid). Overall 2014 has been a watershed year as far as voter turnout is concerned (Figs. 2).

Literacy and education

It is conceivable that a well-educated gentry of a nation would not only make good choices in terms of votes, but also throw in sound and socially acceptable candidates into the race. Yet, we live in times where the ‘tyranny of the majority’ that John Stuart Mill cautioned against in the nineteenth century, has become reality. A system of plural voting, as proposed by Mill, where voters would need to acquire a basic minimum education to secure the right to vote can surely become a reality in modern India, adding to the aspirational quality of literacy, providing a sharp nudge to the upward movement of the literacy rate, which stood at 74 per cent in 2011, with a sluggish 10  per cent gain per decade since 1981 (Census, 2011). Then again, a weighted system of counting votes, where votes of candidates with higher qualifications would count more, as suggested by Mill, could also be enabled. However, there is the fear amidst scholars that if literacy and level of education is brought into play, the elite would capture the political process resulting in an elitist status-quo, making change near impossible (Bardhan and Mookherjee, 2000). It is, however, necessary to add that literacy, education and access to information are divaricate parameters and the premise that the illiterate select poor candidates may not altogether be true. It may also be true that the highly educated abstain from voting. For example, no discernible pattern emerges when state-wise data for enrollment into higher education is considered against voter turnout. In fact, the correlation between 2014 voter turnout and enrollment in higher education (based on data collected from NITI Aayog for 2012-13), shows a slight negative trend (-0.01), perhaps indicating that the highly educated keep away from voting. It may be noted that higher education enrollment also includes migrant youth which may need to be factored in for further research. When, however secondary and upper secondary data for 2012-13 is correlated against the 2014 voter turnout, an insignificant but positive trend emerges at 0.1 per cent for secondary and at 0.01 for upper secondary. Further probe into disaggregated district-wise data will help ascertain education’s role in voter turnout. Interestingly, Bihar, occupying the lowest rung in almost all enrollment ratios, exhibits a consistent interrelationship with low voter turnout (Table 1). However, the states that show a high voter turnout, do not necessarily occupy the highest end of the enrollment spectrum, indicating a complex reality.

Despite this the role of education in informed decision making cannot be denied. A study by Bhue  et al. (2014), using postal ballot as the determinant, brought about significant revelations about educated voters exhibiting progressive preferences. They found that educated voters’ support for candidates with criminal records of heinous crimes, such as rape, murder, kidnapping, etc, were a significant 46.5 per cent lower as compared to other candidates. Moreover, educated candidates exhibited a lower preference (by 26.5 per cent) for candidates with higher net assets – which may be understood as a proxy for corruption. Educated candidates also showed a marked appreciation for female candidates – with 12.7 per cent greater support as compared to male candidates. The authors thus conclude that educated voters make better choices. The study however suffers on two counts –one that the analysis relies heavily on a ballot that constitutes a mere 0.4 per cent of total votes, and second there is no factoring in of educated voters (from proxy variables) in the total voter turnout. The fact that education affects positively is a take away from the study, although the actual extent may be contentious.


Economic development is not equitable in India—there exist wide divergences in state-level performance, whether it is increasing competition, reforms such as land and labour laws, rising unemployment, debt and deficit and more—some states perform well while many others lag dismally (CRISIL, 2018). A cursory look at the per capita net state domestic product (PCNSDP) for 2014-15 computed at the base year 2011-12, shows an indicative trend. A simple co-relation between voter turnout and PCNSDP works out to be 0.1 for the country, marking a weak relationship between higher per capita income and voter turnout. Income thus may have little bearing on people’s propensity to cast a vote at the aggregate all India level. However, regional differences exist. For example states such as Bihar, Uttar Pradesh and Madhya Pradesh show the lowest voter turnout and at the same time occupy the lowest rungs in the PCNSDP.  Clearly, the driving need in the people residing in these states is to eke out a living, rather than investing time to stand in a queue and participate in the voting humdrum. Uttar Pradesh (with 95.5 million persons), Bihar (with 35 million persons)  and Madhya Pradesh (with 23 million persons) host India’s largest populations that the erstwhile Planning Commission, now NITI Aayog, has earmarked to be below the poverty line  (Reserve Bank of India, 2013). However, despite the fact that a poor state of economy may exert an influence on the voting behaviour, all states do not conform to this pattern. States in the north-east, especially Assam, have a low per capita income as well as a significant number of people below the poverty line, yet they bear testimony to a high voter turnout. West Bengal too does not fall on the extreme opposite end of the scale, despite showing a very high voter turnout. Perhaps there exists an exogenous link between poverty and various other factors that result in high turnout in certain states—being low on development indices alone does not hold true.

This section, however, would remain incomplete without outlining what scholars have attempted to analyse—economy and the incumbent government. It is an attractive notion to believe that if the economy performed well under the watch of an incumbent government the voters would reward it and vice versa. The prevailing ideas of democratic accountability and retrospective voting, especially in the US, make it pertinent to analyse how economic growth is related to voter turnout and preference (Vaishnav and Swanson, 2015). Voting based on events of the past for the aggregate period from 1980 to 2012 shows no evidence of an overall relationship between growth and electoral performance (ibid.). However, the electoral timing is important as the voter is most likely to be predisposed towards voting for the incumbent party based on the most recently available information, rather than its entire tenure. Therefore a short spurt of growth before an election may be able to exert considerable influence on winning back votes for the incumbent party. Vaishnav and Swanson conclude that there seems to be a shift in voter behaviour where it may appear that ‘good economics make for good politics’.  Another study by Gupta and Pangariya (2014) cites positive associations between economic growth and the prospects of incumbent governments – where higher the rate of growth, larger the percentage of ruling party’s candidates winning seats.Interestingly, Asher and Novosad (2017) have shown that constituencies experience higher economic growth and employment generation under ruling candidates as opposed to being under opposition candidates. Consistent findings are however elusive as the intelligentsia differ on the metrics used to measure the electoral performance. Scholars have in fact, in addition to the common economic indices, used private sector employment, share prices, increased output measured by night lights and more to ferret out possible associations between politics and economy.

Voter Turnout | Gender

The past few decades have shown a continual decrease in the gender gap for female voter turnout. Where gender gap stood at 17 per cent in the 1957 General Elections, by the turn of the century, it had decreased to 8 per cent, further reducing to 1.46 per cent in the 2014 General Elections. This change, starting most significantly in 2008 was witnessed in almost all states in varying degrees. An interesting pattern was witnessed in states where gender gap was earlier narrow—during the period between 2008 and 2018, the voting pattern in various assembly elections had reversed and women were voting in higher numbers than men. Three distinctive groups of states emerge, based on the extent of gender gap—high, moderate and low gender gap states. The states of Bihar, Madhya Pradesh, Assam, Rajasthan and Uttar Pradesh fall under the first category, while the second category includes the states in South India, with the exception of Kerala. West Bengal and the north-eastern states fall under the third category, where the gender gap had not only narrowed, but women had outnumbered men in voting (Kumar and Gupta, 2015).

Class, Caste and Religion

Most of the discourse that relates to class, caste and religion analyses poll results and debates about who voted and for whom. Despite the focus of the paper being a spatial analysis of voter turnout, the digressions into post poll analysis is necessitated by the lack of comprehensive data on the socioeconomic composition of the voter turnout. Sridharan (2014) argues that the growth of the middle class has “changed the bi-polar elite mass social structure” that existed during Independence to a “roughly three-layered elite-middle class-mass social structure”.  In the Indian context he opines that the upper and middle class constitute 47 per cent or nearly half of the population, which he adds are disproportionately upper caste as well. The 2014 elections form the backdrop of the study where he contends that there is a higher turnout amongst the richer classes as compared to the poor. This is consistent with our earlier premise that education extends a positive impact, as well as the premise that poor may find the polling exercise wasteful as opposed to salvaging their livelihood. Interestingly, the author points out that the Hindi belt and western states, where we see a marked increase in voter turnout, may hold attitudes that are anti-minority. However, the polarisation needs to be also perceived in the light of middle and upper classes preference for government expenditure on infrastructure rather than subsidies for the poor (ibid). Class politics is one of the several axes in the context of economic policies of developed nations—still not a norm in India. A couple of elections more may give an idea of
class preferences.

Scholars analysing data on the largest minority community in India, the Muslims, provide yet another perspective on voter turnout. Religious minorities tend to be underrepresented in many democratically elected bodies that Heath et al (2015) identify as supply and demand discriminations. A faulty selection process on the supply side and voter preference for a non-minority candidate, leaving minority success limited to areas of large minority population on the demand side—are issues that govern electoral politics. The implicit assumption, however, is that voters will favour the candidate over the party. Another assumption that dominates the discourse is that Muslims would participate in the elections as a cohesive unit, voting en bloc to influence the outcomes (Alam, 2009). Alam argues that the overall turnout increases substantially when the proportion of Muslim population is well beyond a threshold 30 per cent, yet below the 50 per cent mark. Thus, the population needs to be statistically significant nevertheless needs the demographic balance tipped against the community to favour a high voter turnout. In his study, Alam points out that in eight of the 24 constituencies in 2009 where Muslim population was 30-40 per cent high turnout was caused by competitive communal mobilisation. However, the argument fails to explain the high voter turnout in 11 constituencies in West Bengal and Kerala where overall turnout has been usually high. The author offers an explanation for higher Muslim voter turnout in Assam where he opines the Muslim population has grown large enough to influence or even dominate poll outcomes. In states such as Madhya Pradesh, Rajasthan and Maharashtra the low voter turnout is attributed to the relatively small size of the population (ibid).  Muslim participation thus appears to be largely contextual and voter turnout is influenced by a “complex mix of factors that operate in different spatio-political contexts”(Alam, 2009).

Political awareness

Political awareness is the key to a vibrant democracy, which is not only about a healthy voter turnout, but also about having adequate information about parties, policies, candidates and the election process itself so that voters can make an informed choice. Radio, television and newspapers were known to exert considerable influence in the yesteryears —today it seems that smartphones and the internet are subtly changing the mechanisms of information access.  The Indian Readership Survey 2017 holds certain key statewise trends. The irregular availability of electricity in Uttar Pradesh, Bihar and Jharkhand may be treated as a proxy variable—understood to exert a detrimental effect on access to information—which coincides with poor voter turnout in these states. When assets such as TVs are considered, the poor turnout of Bihar and Uttar Pradesh is again corroborated, with these states occupying the lowest two rungs at 22 and 42 per cent in TV ownership (MRUC, 2017).  As far as fixed-line phones are concerned, it comes as no surprise that Bihar remains at the bottom of this list at 60 per cent tele-density and Uttar Pradesh hangs on only a little above. On the other hand Kerala, with one of the highest voter turnout has a tele-density of 119.35 (TRAI, 2018). As far as the number of mobile subscribers is concerned, Uttar Pradesh presents the largest anomaly topping the list in 2016 with the highest number of mobile users, with Bihar following close behind. This however, may be a function of a heightened need for connectivity, as this area has been historically deprived of infrastructural amenities (OGD
India, 2017).

Print media also has a considerable turnout on election outcomes. Newspapers, for example, are strong determinants as far as choice of political party is concerned. A study (Banerjee and Kumar, 2018) conducted for the 2011 State Assembly elections in Kolkata held that high readership of newspapers with anti-incumbent views had a significant impact on the outcome of the elections.

Another variable that can be used to understand the level of political motivation and awareness is grass root electoral politics—the Panchayat elections. The intensely fought local contests showcase frequent faceoff between candidates. For a Panchayat seat, where votes are in mere thousands, every vote assumes great importance. Candidates and campaigners reach out to the last voter, persuading and urging them to show-up for voting (Banerjee, 2014). Thus everyone has first or at least second hand knowledge of the candidates. A select statewise data of various panchayat elections throw up two observations—first that states show a significantly higher voter turnout in Panchayat elections and second that Panchayat elections are marred by violence. Haryana is a case in point that showed a massive voter turnout of 87 per cent in Panchayat elections (State Election Commission Haryana, 2016), whereas its general elections’ poll turnout hovered around 70 per cent. States such as West Bengal, which otherwise hits a high voter turnout, suffers frequent incidences of violence during Panchayat elections lowering its voter turnout in the last instance to 72 per cent (Business Standard, 2018).

Limitations and way forward

Voter turnout is understandably affected by various factors, some of which have been briefly enumerated here. However, there are a few concerns that need to be flagged before further discussion is enabled. A huge body of data is available along geographical and administrative lines —yet elections are conducted constituency-wise, for which no data spectrum exists. The parliamentary constituencies are not contiguous with administrative or any other units, making it a unique political bastion. Based as they are on population, the constituencies criss-cross districts, tehsils/ taluks/ blocks and so on. There is thus no way to collect direct evidence of how far the political representatives have taken care of those who made them legislators (Alam, 2010) among others. Most scholars thus rely on creating their own socio-economic profiles based on patterns they desire to decipher. Therefore the question of transparency, accountability and academic rigour rises to the fore.

Some opine that building a National Electoral Roll Purification and Authentication Programme (NERPAP), through an error-free, purified and authenticated electoral roll may be the first step to authentic data. Introduced in March 2015, by the Election Commission of India (ECI) the NERPAP programme would link the Electoral Photo ID Card (EPIC) with the Aadhaar data for the purpose of authentication. The voluntary Aadhaar-EPIC link could be fed through the National Voters Services Portal (NVSP) or furnished through SMS/email/submission of copy of documents to Electoral Registration Offices or Booth Level Officers (PIB, 2015). While the programme was in progress, in August 2015, the Supreme Court of India, through an interim directive in the case of Justice K S Puttaswamy vs Union of India, ordered that apart from the use of Aadhaar card for Public Distribution System schemes, no other usage was permitted. Following the interim order, the ECI suspended all activities related to collection of Aadhaar number (ECI, 2015).

It was not until July 2017, that the ECI started filing applications to the Supreme Court for granting it permission to continue with the (voluntary) linking of EPIC and Aadhaar. In March 2018, it was reported that the ECI had switched its position on the linking and had filed a fresh application in the Supreme Court to make linking/seeding compulsory (Financial Express, 2018).

The intent of the ECI behind linking the EPIC number with Aadhaar is clear—controlling electoral malpractices such as the entry of fake or duplicate names on the electoral roll. Right before the NERPAP was officially launched, the then Chief Election Commissioner (CEC) H S Brahma stated to the press that about 85 million names on India’s electoral rolls were incorrect (Hindustan Times, 2015).  CEC OP Rawat, in his statement to the press in January 2018 noted that linking Aadhaar with EPIC would make it possible for migrant workers to cast their votes like members of the armed forces. Further, biometrical identification would help establish the identity of a voter before they enter the polling booth and get access to the EVM. This, however, does not entail any changes in the EVM, which will continue as a standalone device with no connection to any other device or the internet (Sharma, 2018).

There are however certain cautions as well. Data leakage and theft are realties of the modern age—that is likely to hound us extensively in the future as well. We are also aware of how opposition led states are systematically misaligned, with funds being withheld by the ruling party. However, if a situation arises, where citizens are revealed to be anti-incumbent, they may be specifically targeted, making them poorer as a result, as Hseih (2011) point out in a case study of Venezuela.

Another area of study that has largely remained untouched is how the youth are perceiving the political milieu. Young people are not actively voting in the developed world and before long we may see the reflections of this trend here in India too. Being a young nation this could affect the political environment significantly. The 2019 General Elections add an entire generation born at the start of the 21st century, with 133 million first time voters.


Voting is a collective action, where one vote amongst a million holds little relevance. Yet, every vote counts in making the electoral process resoundingly secure. There are many who desist with little motivation to actually make the trip to the voting booth. Corruption and criminal records of candidates dishearten many. Yet, residents of certain states vote passionately and unfailingly while others sway, riding the wave of the day. What makes them vote is hugely contextual —yet some patterns do emerge. Low voter turnout in Bihar and Uttar Pradesh can be easily correlated with most of the parameters outlined above, from developmental constraints to low literacy and low levels of awareness. These states have shown the most consistent outcomes. The states with high voter turnout, with special reference to the states in the East and North-East, provide a varied picture. West Bengal despite being located midway in most indices, hogs the limelight in voter turnout. Most notably Assam, where voter turnout is extremely high, more so for women than men, despite having one of the lowest access to development, education, assets in terms of TV and phone, low PCNSDP and more, presents an interesting anomaly to the pattern. Although there is a need to research further these unique voter turnout strongholds, some preliminary observations about West Bengal’s renaissance awakening and Assam’s political volatility since the 1980s students movement may be made. The political commitment perhaps helped to drive the high voter turnout. We can hope that the 2019 elections will provide finely graded voter turnout data, enabling further reflections with élan.


Alam M.S., 2010. On Matching Census Tracts and Electoral Boundaries: The Bottom-up Aggregation Approach, Economic and Political Weekly, 35(34): 64-72.

_________, 2009. Whither Muslim Politics? Economic and Political Weekly, 39(26).

Asher S. and P. Novosad, Politics and Local Economic Growth, American Economic Journal: Applied Economics, 9(1): 229-273.

Banerjee M., 2014. Why India Votes?New Delhi: Routledge.

Bardhan and Mookherjee, 2000. Capture and Governance at Local and National Levels, American Ecomomic Review, 90(2): 135-139.

Bhue G.S., M.R. Rao, K.V. Subramanian, P. Tantri, 2014. Does education affect voting preferences? Evidence from voting records of Indian elections, Indian Statistical Institute.

Business Standard, 2018. West Bengal panchayat polls: 72% turnout; violence claims 12 lives, Business Standard, May 14.

Election Commission of India, 2014. State wise Voter Turnout, Available at:

Election Commission Of India, 2018. Status paper on electronic voting machine: Election Commission of India. Available at:

Financial Express, 2018. Aadhaar linking news update: Election Commission appeals for mandatory seeding of 12-digit UID with Voter-ID, Financial Express, March 11.

Ford, 2014.India’s democracy runs on 1.7 million electronic voting machines, Quartz, April 16.

Gupta P. and A. Panagariya, 2014. Growth and Election Outcomes in a Developing Country, Economics and Politics, 26(2): 332-354.

Heath O., G. Verniers and S. Kumar, 2014. Do Muslim Voters prefer Muslim Candidates? Co-religiosity and voting behaviour in India, Electoral Studies, 38.

Hseih C., E. Miguel, D. Ortega and F. Rodriguez, 2011. The Price of Political Opposition: Evidence from Venezuela’s Maisanta, American Economic Journal: Applied Economics, 3(2): 196-214.

HT Correspondent, 2015. 85 million fake or duplicate names on electoral rolls: EC, Hindustan Times, February 28.

Media Research Users Council, 2017. IRS 2017: Key Trends, Available at:

National Journal Leadership council, 2016. 2016 voter turnout and demographic breakdown. Available at:

Open Government Data (OGD) Platform India, 2017. Mobile Telephone Subscribers as on 29.02.2016, March 22.

Press Information Bureau, 2015. National Electoral Roll Purification and Authentication Programme (NERPAP), 2015 – Collection of Aadhaar number from the applicants– Regarding, Election Commission, May 29.

Ramani S., The “High Turnout” in ongoing Lok Sabha polls – What could explain it?Economic and Political Weekly, 49(18).

Reserve Bank of India, 2013. Number and Percentage of Population Below Poverty Line, Handbook of Statistics of Indian Economy.

Sharma S.N., 2018. I support linking Aadhaar to voters’ identity: New CEC Om Prakash Rawat, The Econominc Times, January 22.

Sridharan E., 2014. Class Voting in the 2014 Lok Sabha Elections: The Growing Size and Importance of the Middle Class, Economic and Political Weekly, 49(39).

State Election Commission Haryana, 2016. Report on General Elections to the Panchayati Raj Institution in Haryana, Available at:

Vaishnav M. and R. Swanson, 2015. Does Good Economics make for Good Politics? India Review, 14(3): 279-311.

Leave a Reply

Your email address will not be published. Required fields are marked *