THE now historic 2013 Delhi Assembly election introduced the state, and the country, to a new party – the Aam Aadmi Party – brought in by the energy of hundreds and thousands of Delhiites taking to the streets to assert their footprint on the collective vision of the capital’s new governance regime. This democratic exercise brought with it highly publicized figures of a strong voter turnout – a so-called long-awaited ‘awakening of a political class.’ Yet, amidst this narrative of triumph, another story slipped by without much fanfare – a Delhi whose democratic system was quietly undermined due to the improper maintenance of urban electoral rolls.

At its very essence, the voter list is the key determinant of who gets to cast a vote on election day; serious errors in the lists can have a massive impact on the outcome, including casting doubt whether the elections can truly be considered free and fair. As elections unfold, debate across the country is focused around opinion polls, on expected vote share, funding of elections, criminal records of candidates, ‘get the vote out’ campaigns and so on. But somewhat intriguingly, of all electoral reform agendas in India, the quality of electoral lists is arguably the most neglected. Although the Election Commission is a credible institution that has done an admirable job of conducting clean elections, it is not equipped to meet the logistical challenges it is confronted with, particularly across India’s growing cities and towns.

The 2013 Delhi Assembly elections are a stark example of this ‘tale of two cities’, in which the quality of the voter list plays a central role in overriding all other electoral issues and outcomes that invariably take centre stage.

The Janaagraha Centre for Citizenship and Democracy (Janaagraha), a Bangalore based NGO focused on improving the quality of life in India’s cities and towns, began to work on the quality of electoral lists a decade ago after a pilot in two Bangalore Assembly Constituencies (ACs) revealed that close to half the names in the voter list were erroneous and required deletion – either because the voter had died, shifted, or could not be found. Further work on this issue demonstrated that poor maintenance of the electoral list is a result of migration and scale, making the task a herculean logistical challenge.

In order to ensure sustainable advocacy for change, Janaagraha helped create a Proper Urban Electoral (PURE) process using technology and GIS mapping, neighbourhood level feet-on-street for year-round updating, improved collection of data and its transparent use in the public domain, created more channels for citizens to update their information, and tightened vigilance on election day to ensure that errors in the list are not misused as ‘phantom votes’. Based on our ability to access data, the Election Commission of India (EC) entered into a formal MoU with Janaagraha in 2010. Currently we are implementing our system in 27 Assembly Constituencies (ACs) of Bangalore. As part of the PURE effort, Janaagraha began conducting rigorous systematic surveys in cities across India to provide never-before collected data on the level of error existing in voter lists.¹

The Delhi PURE List survey, conducted in the run-up to the assembly election, found that close to 20% of the names appearing on Delhi’s voter lists could have serious errors, requiring these names to be deleted either because the voter had died, shifted, or could not be found. In effect, therefore, these names were ‘Phantom Voters’. Importantly, the study was conducted post the EC clean-up of Delhi, which reportedly removed 14 lakh names (see table I for AC-wise data).

What could these figures mean in actual numbers? Out of the 12.3 million registered voters on Delhi’s list, more than 23 lakh require removal. Although the error rates fluctuate across the AC’s surveyed, even the one boasting of the lowest error rate (Ambedkar Nagar at 8%) still adds up to around 10,000 names to be deleted. In constituencies like Najafgarh and Burari, 24% of the names on the list need to be deleted – tallying over 45,000 and 55,000 voter names respectively.

Critically, the implication of these results stretch far beyond an administrative headache. In the seven assembly constituencies in which the PURE List study was undertaken, the margin of victory in 2008 in six constituencies was less than the number of names that needed to be deleted from the voter list. In Burari AC, when the number of names to be deleted was over 45,000, the margin of victory was 4990 votes. In Shahadra AC the margin of victory was 1536 votes; the number of names to be deleted was over 25,000 (see Table II for breakdown of margin of victory vs. names to be deleted).

Beyond the issue of neglected deletions, a second prominent issue arises from the PURE list studies: mass additions to the voter list made in a short time-frame close to elections. In the context of Delhi, data on mass additions was not available, as it required voter list data to be collected over a period of years. No wonder that this issue often disappears from the public domain shortly after an electoral exercise. Janaagraha only recently introduced the PURE process to Delhi. Data collected and aggregated over a period of years from Bangalore will therefore be used to illustrate this issue.

On comparing changes in the voter lists of Bangalore Urban between the 2008 election and the recent 2014 Lok Sabha vote, taking into account all additions and deletions made within this time frame, the total number of voters went up by a mere 80,683 names in this five year period.² However, between the 2013 assembly elections and 2014 parliamentary elections, the total number of registered voters increased by 5,29,995 in barely nine months.³ It is worth examining how the Bangalore list increased by over 500,000 in this period.

 

Bulk additions such as these create a major obstacle to ensuring due process, as in cities, and particularly at the critical time of elections, the EC does not have its own machinery to rely on. It is the state machinery (e.g. teachers and postal workers) who are tasked with full-time non-election duties. For an already stretched operation, bulk additions create an improbable task for ensuring proper verification.

Both mass additions and neglected deletions within the list create a huge pool of potential ‘phantom voters’, opening the door to fraud voting. In April 2013, the EC of Delhi uncovered 13 lakh bogus voters and inflated voter numbers in 20,000 households.⁴ Even the Supreme Court of India recently acknowledged gaps in checks and balances in voting processes, stating that the ‘none of the above (NOTA)’ voting option would help prevent fraudulent proxy voting that currently takes place.⁵ This year’s Lok Sabha elections ended with an alleged case of false registrations in Tamil Nadu.⁶

 

As the recent Delhi election clearly demonstrated, elections are increasingly turning into closer contests between a larger pool of candidates. As margins of victory become razor thin (down to a few hundred votes in several Delhi ACs) even a small percentage of ‘phantom voters’ can significantly alter the electoral outcome.

Further, both the central voter list issues – bulk additions and few deletions – result in an understated voter turnout percentage in Indian cities. In other words, when election results are reported, the voter turnout is under-reported, since the names to be deleted are not taken out of the total number of registered voters, even though they are no longer valid.

In Delhi, despite the focus on reporting voter turnout figures, legitimate data to compute voter turnout percentages did not exist. As the table below demonstrates, if the neglected deletions found in the PURE list Delhi study were taken out of the voter lists, the actual voter turnout percentage would be substantially higher.

 

The data from Bangalore provides further corroboration. In Shanti Nagar AC, where Janaagraha helps maintain a clear voter list year-round, between 2013-2014 the voter list showed an increase of 20,306 names and 6,914 more people turning out to vote.⁷ Yet, a reduction in voter turnout percentage was reported. Critically, as the data showed, this was not because fewer people voted, but because the overall number of names on the list was over-stated. Low voter turnout in Indian cities could well be an urban legend, with the villain being the voter list.

 

Importantly, the two major error types highlighted above – deletions and mass additions – are errors of ‘inclusion’, i.e., names that are included in the list but require further verification. However, it should be noted, there are also errors of ‘exclusion’, i.e., eligible voters whose names should appear on the list but do not. For the Delhi study, the focus was on gathering data regarding errors of inclusion, as these were not part of the discourse on voter hygiene and electoral processes. However, we have also collected data on errors of exclusion in past and current studies.

Briefly, we found that attempts to register as a voter do not always succeed. In a Bangalore PURE list study conducted in 2013, we randomly selected people from the voter list (same method as the Delhi study), and additionally, close to 2,000 people per AC by household (thus accounting for rates of omission of people who should have been on the list but were not).

In a constituency where we had implemented the PURE process, 29% of persons were not registered to vote, yet 23% of them had sent in registration forms (with 17% attempting to register more than once). In a control AC, 37% of persons were not registered to vote, yet 58% of them had sent in registration forms (with 41% attempting to register more than once).

Therefore, if similar rates of exclusion in the voter list were to be inferred in the context of Delhi, then the close to 20% error rate, based solely on errors of inclusion, would soar.

 

Currently, we are undertaking the PURE List survey in 21 major cities and towns across the country, collecting data on errors of both inclusion and exclusion, and whether these errors are correlated to variables such as gender, religion, and socio-economic status. Importantly, existing ethnographic research provides initial data suggesting that citizens in lower socioeconomic categories, particularly those living in Delhi’s slums and unauthorized colonies, face higher rates of errors of exclusion.⁸

With such major implications, how has this issue gone unnoticed? Despite the processes that the EC has mandated to ensure accurate lists, a dynamic and mobile urban citizenry is overwhelming an electoral system that was designed for a rural and fairly static demographic. The ECs current infrastructure and processes are not sufficient to meet the challenge. As cities across the country grow ever faster, it is incredibly difficult for the EC to micro-manage the last mile of every process when resources required are invariably stretched year on year.

Electoral list reform is not just an issue that individuals and institutions outside of the EC point to, but one which many who have worked within the EC also recognize. Recently, former Chief Election Commissioner (CEC) N. Gopalaswami stated that if the EC is to reach its goals, a massive injection of human and financial resource and continuous maintenance of the electoral lists is required.⁹ Note that former CEC, T.S. Krishnamurthy, sits on Janaagraha’s PURE advisory board, in part because of his experience in dealing with these issues.

 

Electoral list maintenance is currently characterized by poor mapping of polling booth boundaries, incomplete addresses and other details of individuals, lack of a sufficiently sophisticated, integrated and open database, temporary staffing and poor incentivization for updating voter lists. Reforms targeted at precisely these five factors are needed to establish the integrity of voter lists, especially in Indian cities where the flux in voting populations is high and rising.

A new system that ensures continuous maintenance and full transparency of voter lists is urgently required. It must involve a much greater role for technology, which could be used to ensure that lists are updated on a continuous basis, mapped onto GIS-based maps, available for public scrutiny and comment, and with more channels open for citizens to update their names as their status changes. The creation of aggregate databases is required, where voter lists are compiled at the AC, PC, and city level (which they are currently not), and where data from previous years rather than only recent data in a disaggregated form is kept open to the public. Enhancing the capacity and resources of the EC should also be a top priority. This would, after all, be an investment to strengthen Indian democracy and possibly call the bluff of urban apathy.

 

Until these processes come into place, in order to ensure the accuracy of current processes the EC should consider disclosing more detailed data on voting, such as booth-wise polling turnout and comparative data on votes cast vs names in the voter lists. Moreover, in the short-term, there is also a greater need for grassroots engagement with communities, since the validation of a name requires not only feet-on-street, but confirmation of identity and residence, which is often not possible due to the quality of information on the lists.

Reforms however will begin only with an acknowledgement of their need. This requires collection of more data on voter list hygiene from cities and towns across the country, as well as research on whether the process of enfranchisement (additions to the list) or disenfranchisement (deletions from the list) are characterized by other compounding factors. This opens up a critical role for both researchers as well as media.

 

The poor quality of voter lists has a direct impact on the quality of Indian democracy itself due to potential disenfranchisement, fraud, and impact on electoral outcomes. The data arising from Janaagraha’s PURE List Delhi study, collected at the peak of an electoral exercise, widely considered one of the most important elections in recent times, creates a pressing call to action. The 2014 Lok Sabha elections further boosted this call, with unjustified deletions of voter names in Pune, Mumbai, Nagpur and Varanasi and the consequent public interest litigation in the Bombay High Court.¹⁰

Reflecting on the PURE List Delhi data, particularly in the wake of the Lok Sabha election, we need to pause for a few months and introspect. As election talk dies down to make room for continuing discussions about what the shape of a new government for the city (and indeed for India) may be like, and about the impending policy changes, it is critical that we also hear the warning bells, which were briefly headline news. If we fail to do so, we may be re-reading the same headlines come the next elections.

The work presented marks just a beginning. More data is required from a wider set of locations to permit comparative analyses. Moreover, data on what correlations, if any, exist in regard to the presence of errors needs to be probed. For example, whether certain groups of people are systematically affected by errors of inclusion or exclusion is of particular concern. Not only would this flesh out a more robust understanding of the problem, but a greater move towards uncovering the root causes and future solutions.

Researchers using both quantitative and qualitative methods should network to produce data-sets that can speak to one another, thus making more efficient use of resources and skill-sets. This data should then be taken up by thought leaders across the spectrum in order to bring about the necessary policy shifts required to ensure robust and legitimate enfranchisement and electoral outcomes.

 

Electoral list hygiene and its intimate connection with the quality of democracy is not a new issue, nor is it India specific. In the last five years, examples of fraudulent voting using erroneous names on the voter list have been documented in large democracies around the world such as the United Kingdom, Israel, Malaysia, and Zimbabwe.¹¹ Further, the exclusion of certain groups of people from the right to vote is also a well documented and global phenomena. A recent example is the worldwide debate that was sparked off when the 2000 United States election saw mass exclusion of African-Americans, as well as those with prison records, from voting in certain areas.¹² Major projects are underway to remedy errors in list hygiene in various countries. The Pew Charitable Trust is working on a massive database mining tool that allows states and the federal government to cross-verify names before deletion, among other things.¹³ Moreover, in various countries, researcher are conducting rigorous studies to understand the scale and causes of disenfranchisement and voter registration processes.¹⁴

As the Delhi case serves to highlight, we are left with serious concerns about a lack of due process in the maintenance of India’s voter lists, often a side of the electoral story that is never discussed. More data, across more cities, and over longer periods of time is required to understand the nature and magnitude of this critical issue and to link it to key questions regarding the nature of current Indian democratic processes. Critically, with Delhi’s current government in flux, and a likely return to the polls soon, this tale of two cities should not be dismissed.

 

Footnotes:

1. Survey methodology: Janaagraha designed the study and questionnaire and survey firm TNS Global carried out the study. A total of 2,847 respondents were selected from 238 polling booths spread across 7 Assembly Constituencies (ACs), giving a 95% confidence level and a confidence interval of +/- 1.8% at the city-level. The Assembly Constituencies (ACs) were chosen using semi-purposive stratified random sampling, stratifying for geographic spread. Within each AC, 34 Polling Parts were randomly selected. From each of the sampled Polling Parts, a total of 10 respondents were randomly selected. Substitute respondents (who were also randomly selected) were only used if the original respondent’s address could not be found after three serious attempts. These ‘not founds’ were included in the final sample and marked as error type ‘not found’. All interviews were conducted face-to-face using a standard structured questionnaire. If the respondent who corresponded to the voter ID was not at home, a family member above the age of 18 who was aware of the respondent’s voter details (spelling of name, whether they were dead, etc.) was surveyed. If there were no family members or they were not aware of the details, three attempts were made to contact the respondent. If the respondent could not be contacted a substitute respondent was selected. The fieldwork was conducted between 5 October and 31 October 2013.

2. These figures are drawn from the Electoral Rolls spanning the years 2008-2014, accessed from the website of the Chief Electoral Officer, Election Commission of Karnataka. http://ceokarnataka.kar.nic.in/ElectionFinalroll2014/Dist_List.aspx Data for each Assembly Consistency was compiled in-house to arrive at aggregate figures for Bangalore urban across this time span.

3. Ibid.

4. Vikram Kumar, ‘Delhi hit by massive poll scam: Election Commission unearths 13 lakh bogus voters and over 80,000 valid voter cards for dead people’ (accessed on 13 August 2013) http://www.dailymail.co.uk/indiahome/indianews/article-2392340/Delhi-hit-massive-poll-scam-Election-commission-unearths-13-lakh-bogus-voters-80-000-valid-voter-cards-dead-people.html#ixzz3 Cr MLOc23

5. Supreme Court of India, Writ Petition (Civil) No. 161 of 2004; People’s Union for Civil Liberties & Anr vs. Union of India & Anr, New Delhi, 27 September 2013; Economic Times, Supreme Court allows voters to reject all candidates in elections (accessed on 27 September 2013) http://articles. economic times. indiatimes.com/2013-09-27/news/42463769_1_negative-voting-clean-candidates-apex-court

6. The Hindu, ‘TN residents enrolled as voters in Kanakapura’ by Shivakumar: BJP (accessed on 26 April 2014) (accessed on 13 August 2013)

7. Op. cit., fn. 2.

8. Manisha Priyam, ‘Pipe-Dreams: Unauthorized Lives and Voter Deliberations in Delhi’s Sangam Vihar’, presentation made at second EECURI Network Seminar, European Social Research Council and Janaagraha, Bangalore, 3-4 January 2014.

9. N. Gopalaswami, ‘The road ahead for the Election Commission’, The Hindu (accessed on 31 May 2014) www.thehindu.com/todays-paper-/tp-opinion/the-road-ahead-for-the-election-commission/article 6068322.ece

10. The Economic Times, ‘20 lakh voters enrolled in Maharashtra: Election Commission’ (accessed on 4 September 2014) www.articles. economictimes.indiatimes.com/2014-09-09.news.53563633_1_election-commission-summary-revision-20-lakh-voters

11. Sam Greenhill and Tim Shipman, ‘Postal vote fraud: 50 criminal inquiries nationwide amid fears bogus voters could swing election’, Daily Mail, UK, (accessed on 4 May 2010) www.dailymailco.uk/news/election/article-1271457/General-Election-2010-Postal-Vote-Fraud-Admid-fears-bogus-voters-swing-election.html; Melanie Lidman, ‘Israel election marred by 400 cases of voter fraud’, Jerusalem Post (accessed on 22 January 2013) www.jpost.com/National-news/israel-election-marred-by-400-cases-of-voter-fraud; Azeem Ibrahim, ‘Election fraud in Malaysia’, The World Post (accessed on 7 May 2013) www.huffingtonpost.com/azeem-ibrahim/malaysia-election-fraud_b_3211954.html; David Smith, ‘UK and US concern over Mugabe’s Zimbabwe election win’, The Guardian (accessed on 3 August 2013) www. theguardian.com/word/aug/03/zimbabwe-morgan-tsvangirai

12. Erika Wood, ‘Florida: how soon we forget’, The New York Times (accessed on 5 April 2012) www.campaignstops.blogs. nytimes. com/2012/04/05/florida-how-quickly-we-forget.;Michael McLaughlin, ‘Felon voting laws disenfranchise 5.85 million Americans with criminal records: the sentencing project’, The Huffington Post (accessed on July 12, 2012) www.huffington post.com/2012/07/12/felon-voting-laws-disenfranchise-sentencing-project_n_1665860.html

13. The Pew Charitable Trusts, Electronic Registration Information Center (ERIC): frequently asked questions (accessed on 4 April 2014) www.pewtrusts.org/en/research-and-analysis/q-and-a/2012/11/2/electronic-registration-information-center-eric-frequently-asked-questions

14. For example, see Vincent Pons et. al., ‘Voter registration costs and disenfranchisement: experimental evidence from France’, Job Market Paper, January 2014. www. economics.mit.edu/grad/vpons/papers