Abusaleh Shariff was member secretary of the Rajinder Sachar Committee, which in 2005 was given the task of studying the social, economic and educational conditions of the Muslim community in India. The committee’s report, submitted a year later, was praised for conclusively showing that Muslims lagged most other social groups in the country.

After stepping down as chief economist at the non-profit National Committee for Applied Economic Research in 2012, Shariff started his own non-governmental organisation, the Centre for Research and Debates in Development Policy, New Delhi. He now divides his time between India and the United States, where he is chief scholar at the US-India Policy Institute in Washington.

In the months leading to the Karnataka Assembly elections slated for May 12, Shariff and his associate Khalid Saifullah have been studying the state’s electoral rolls. Their aim was to track the level of Muslim participation in the political process. Their findings have led them to conclude that nearly 20% of Muslim adults in Karnataka are missing from the voters list. After going through the findings of his study, Scroll.in asked Shariff to compare the exclusion of Muslims with that of other social groups. He prepared comparative tables, which constitute the basis for this interview that involved multiple conversations over phone and email exchanges.

Edited excerpts from the interview:

Abusaleh Shariff says his representations to the Karnataka government and the Election Commission on the missing Muslim voters went unanswered.

What prompted you to study Karnataka’s electoral rolls?
Over the last two years, I have been looking at the issue of citizenship pertaining to Muslims, particularly in the context of Assam, where the government is trying to identify citizens from those who are not. I then went to states like Maharashtra, Uttar Pradesh and Karnataka. I realised that the number of Muslim voters has declined over the years. This realisation became my motivation to study Karnataka’s electoral rolls.

Was this realisation just a feeling or based on concrete data?
It was a feeling that Muslims are slowly withdrawing from electoral politics and that their voting behaviour has changed from the past. With the Karnataka elections around the corner, I wanted to tie into the participation of Muslims in the electoral process a little more deeply.

Were you given access to the electoral rolls in Karnataka?
My team extracted the data from the official website of the Karnataka Election Commission. It is an open resource. The extraction was done in February. Karnataka then had 4.98 million voters. Subsequently, in April, the Karnataka Election Office issued a press release saying it had enrolled 1.3 million additional voters during a three-day mass enrolment programme. So, Karnataka now has 6.2 million voters.

What was your principal finding?
A large number of Karnataka’s adults are not on its voters list. This number is particularly high for Muslims – 20% or 13 lakh Muslim adults in Karnataka will not vote in the May 12 elections only because they are not on the electoral rolls. In comparison, 12.3% or 53.2 lakh non-Muslims, defined in our study as “All-Others”, will not be able to vote for the same reason. This finding should be seen in the context that Muslims constitute just 13% of Karnataka’s population.

Why have you chosen to compare Muslims with All-Others, not with, say, Hindus or Scheduled Castes?
It is easy to identify Muslims and create a list of voters from the community because of their unique names. By contrast, it is difficult to create a list of Scheduled Castes or upper castes or Other Backward Classes. This is because it is not always possible to identify a person’s caste by his or her name. Likewise, non-Muslim minorities have names from which you cannot always make out their religious affiliations. So, for the sake of comparison, we created the All-Others category, which comprises Hindus plus non-Muslim religious minorities.

How do you figure out missing voters from the voters list? It sounds like a contradiction.
First, we created a data bank of Muslim voters. It was a list of Muslim households by voters. Our project was designed to study the presence or absence of Muslim voters from the electoral rolls. Since Scroll.in wanted the results for Muslims to be compared with that of one other group, we created a category of All-Others for a list of 25 constituencies. In each of these 25 constituencies, at least 10% of its total households belong to the Muslim community. We then prepared a list of households by voters for both Muslims and All-Others for the 25 constituencies.

What exactly do you mean by households by voters?
We expect each household to have at least two voters – a husband and a wife. In some households with grown-up children, you will have the husband, the wife and their grown-up but unmarried children. There would also be instances of joint families, three generations living together.

But the minimum votes you expect from a household are two – of the husband and wife. But in any society, there are people who also live alone. In Karnataka, 4.7% of all households have just one person. These are known as one-person households.

Is this the Census of India’s figure?
Yes. So, by using the list of voters and creating the distribution of households…

How did you create the distribution of households?
It is very easy. Each voter has an address. If there are three voters in a house, then that address will be repeated three times. But if an address is not repeated, it means there is only one voter there. If the address is repeated two times, it means there are two voters. These are real people with addresses, where we can go and shake their hands.

We segregated all one-voter households from those with two or more voters. In all 25 constituencies, the number of one-voter households for both Muslims and All-Others outstrips, by a margin, the Census’ figure of 4.7% single-person households in Karnataka. From this, we concluded that households in excess of 4.7% in a given constituency have at least one adult who is not on the voter’s list.

However, the incidence of exclusion among Muslims is nearly 24 percentage points more than among All-Others. This you can see from Table 1 and Table 2.

Table 1: Identification of households with one or more unregistered ‘Muslim’ adults in specified constituencies in Karnataka as on March 2018

Table 2: Identification of households with one or more unregistered ‘All-Other’ adults in specified constituencies in Karnataka as on March 2018

For readers who are not adept at reading data or do not have a background in statistics, can you explain the tables by parsing, say, the first entry – that of Haveri Assembly constituency, which is a reserved seat?
Table 1 compares the statistical profile of all households, regardless of the social identity of their inhabitants, in Haveri with that of Muslim households there. Let me first explain the set of data under the caption “All Households”. As per the voters list , there are 47,451 households [Column A]. The Census says 4.7% of all households on average in Karnataka comprise a single resident. This means that we should expect 2,234 single-person households [Column B]. According to the voters list, however, there are 17,118 single-voter households [Column C]. In comparison with the Census data of 4.7%, we have an excess of 14,884 single-voter households [Column D]. This means that at least one member in 14,884 households in Haveri is missing from its electoral roll.

Who are they? What is the social identity of these missing voters?
For an answer to these questions, look at the data in Table 1 under the Caption “Muslim Households”. As per the voters list, there are 6,140 Muslim households [Column E] in Haveri. The expected single-person Muslim households, calculated on the basis of the Census figure of 4.7%, is 289 [Column F]. But the voters list shows there are 3,662 single-voter Muslim households [Column G] in Haveri. This means there are 3,375 single-voter Muslim households more than what should have been [Column H]. Or, to put it another way, 55% of all Muslim households in Haveri have just one voter [Column I]. This means at least one adult in 3,375 households has not been registered as a voter.

You are multiplying 3,375 by one. What is the logic?
Yes. We have assumed that in 3,375 households, there must be at least two votes, most likely one of a man [husband] and a woman [wife]. This is the minimum number of voters per household. So one is on the list, the other is not.

I suppose you have followed the same steps to count missing voters in the All-Others category, which includes everyone but Muslims?
Yes, Table 2 compares all households in Haveri with households in the All-Others category in the same constituency. Under the caption “All-Others”, you can see that there are 11,508 single-voter households [Column H], as against the expected 1,946 households [Column F]. In other words, nearly 28% of All-Other households [Column I] consist of just one voter. Or, to put in another way, at least 11,508 voters in the All-Others category are missing from the voter’s list.

Table 3: Differential additional households with unregistered adult between ‘Muslim’ and ‘All-Others’ groups

 Assembly Name % Additional Muslim Voter HH  % Additional All Others Voter HH Difference
    Haveri 55.0 27.8  27.2
27.3 27
Navalgund 59.9 33.1 26.8
Ron 56.4
30.1 26.3
Hadagali 59.2 33.3 25.9
Kundgol 52.3 27.0 25.3
Byadgi 49.6 24.4
Harapanahalli 50.6 25.5 25.1
Kittur 57.4 32.7 24.7
38.5 24.3
Devara Hippargi 51.0 26.8 24.2
Mudhol 53.8 29.7 24.1
Devanahalli 50.1 26.1 24.0
Sakaleshpur 48.9 25.1 23.8
Basavana Bagevadi 56.1 33.1 23.0
Kampli 56.4 33.4 23.0
Bangarapet 49.8 27.2 22.6
Sringeri 47.4 24.9 22.5
Saundatti 54.9 32.7 22.2
Kushtagi 55.4 33.5 21.9
Karwar 63.3 41.5 21.8
Ranebennur 55.4 33.7 21.7
Hirekerur  47.3 25.8 21.5
Arsikere 45.6 24.2 21.3
59.7 38.4 21.4
Total 54.3 30.6 23.7

Table 3 [above] shows the difference between missing Muslim voters and missing All-Others voters. For the 25 constituencies, 54.3% of all Muslim households have just one voter; it is 30.6% for All-Others. Thus, single-voter households among Muslims is nearly 24 percentage points higher than among All-Others.

Table 4: Estimates of unregistered / excluded adults from voters list in Karnataka as on March 2018

All Voters Muslims All-Others

Total voters reported by the chief election officer, in lakhs *

(Share of Muslims in Karnataka population is 13%)

498 64.7 433.3

CRDDP estimated net single person voter households with one or more unregistered adults, in lakhs

66.2 13 53.2

CRDDP estimated net single person voter households as a percentage to total voters (%)

13.3% 20% 12.3%
* 4.7% from Census 2011; CRDDP is the Centre for Research and Debates in Development Policy

According to you, 54.3% of Muslim households and 30.6% of All-Others in Karnataka are single-voter. But the Census says only 4.7% of all households in Karnataka are single-person. What then is the basis of calculation in Table 4 [above], which shows that only 20% of the Muslim electorate and 12.3% of All-Others have been excluded from the voter’s list?
You must remember that there are different types of households in Karnataka. You have joint households. There are households where unmarried adults live under the same roof. Our data for Karnataka shows that there are 66.2 lakh households in Karnataka, of which 13 lakhs are Muslim and the remaining 53.2 lakh All-Others. These absolute numbers are derived from the Centre for Research and Debates in Development Policy’s basic data files from which only total counts can be achieved. Using the absolute population data from the chief electoral officer and single-person household estimates from our data, we calculated that 20% of Muslim adults and 12.3% of All-Others have not been registered as voters.

But you must remember that 20% and 12.3% represent minimum exclusion. It also follows from this that our margin of error can only be negligible.

Can we, on a sample size of 25 constituencies, extrapolate the results for the entire state?
Twenty-five constituencies constitute 11% of the 224 constituencies in Karnataka. Any statistician will tell you that an 11% sample is good enough to understand an all-pervasive event in the universe of 224 constituencies.

Isn’t the overlaying of 4.7% of single households [from the Census] on each Assembly constituency problematic?
These shares were estimated at the district-level across Karnataka. By the way, variations in these shares range between 4% and 5% even for urban concentrations. The state average of 4.7% is, therefore, a good benchmark to use to make adjustments. Given the fairly low rates of single-person households, this factor alone is unlikely to contribute to errors, if any, in our study.

Couldn’t there be problems with the data you extracted from the Election Commission’s website?
I have 40 years of experience in data collection and data mining. All the protocols through which any data has to pass were followed, including cleaning and setting. The procedures we followed were labour-intensive and often, meticulous matching of individual cases became necessary.

The data uploaded by the chief election officer of Karnataka on its website in February was not in a usable format. We used advanced computer programming to convert the data from the pdf format to the word and excel formats. Also, the data was in Kannada. Through programming, we converted it into a readable English format.

Since you relied on residential addresses to identify single-voter households, it is possible that even things like two spaces between words could throw up different results and skew the results.
Our data extraction was at the booth-level and, consequently, matching voters with household numbers did not pose much of a problem.

What was the procedure adopted to identify Muslim names?
The process of isolating Muslim names was done through a database of all possible Muslims names and their variations in spelling, abbreviation and so on. In the chapter “Access to Credit” in the Sachar Committee report, this was the procedure adopted to extract data from nationalised banks with the Reserve Bank of India’s support. So, it is not that we do not have skills in using this methodology.

Do you have the same kind of table for each of Karnataka’s 224 constituencies as you have for the 25 that you sent across to Scroll.in?
Till date, we have prepared tables for single-voter Muslim households for 100 constituencies. As I said earlier, we wanted to study the extent to which Muslims have been excluded from the voters list. But the list for All-Others is confined to 25 constituencies, specifically prepared for Scroll.in. Obviously, this list can be expanded. But it is labour-intensive and requires funds, which we do not have. We went ahead and prepared the list for All-Others because we thought it was a good idea to compare the exclusion of Muslims with All-Others.

How do you read your own findings?
Obviously, I wish to know the social identity of the missing voters in the All-Others category. My sense is that it would be marginalised groups like Dalits. I say this on the basis of the Muslim community being a marginalised group – as the Sachar Committee report showed – and nearly 20% of its adults in Karnataka not being on the electoral rolls. Muslims constitute just 13% of the state’s population – of that 13%, a section would be below the age of 18. With 20% of Muslim adults not even on the voters list, you can very well see the level of their disenfranchisement is unconscionably high.

I think it is important to point out that our data also shows that a large chunk of Muslim areas in certain constituencies have been left out of the voters list. Let me give you an example – there is a slum area in Bengaluru with 23,000 single-voter Muslim households, or more than 60% of all households in that area.

Did you bring the exclusion of Muslims and others from the voters list to the Karnataka Election Commission’s notice?
We did bring the issue of exclusion of Muslims from the voters list to the notice of the state’s minority commission and the Karnataka government secretary who liaises with the Election Commission. We also wrote to the Election Commission of India and the chief election officer in Karnataka. While there was a mass enrolment programme that reportedly added 1.3 million voters to the rolls, it cannot be that everyone was included.

The exclusion of adults, regardless of their caste and religion affiliation, from the voters list is a serious issue. It is particularly so for marginalised groups that look upon democracy as a process through which they can alter their lives and transform the social landscape in which they are embedded.

What could be the reasons for only one person in a household enrolling as a voter?
It could be a demand and supply kind of problem. For instance, over the years, Muslims have come to feel that it does not matter whether they vote. For instance, the Bharatiya Janata Party has not fielded a single Muslim candidate in many recent elections. It also says it does not want Muslim votes. This kind of discourse discourages or holds back Muslims from enrolling themselves in the voters list. This could very well be the psychology of other marginalised groups too.

Second, those entrusted with preparing the voters list simply go to a chauraha [crossroads], point at homes and ask who their owners are. Normally, the name of the person heading the household is noted and the other members left out. They do not conduct a house-by-house survey. Guiding the “enrolment officer” is often a person from a dominant social group.

Third, there could be a systemic bias against Muslims, a bias so strong that leads to the exclusion of an entire locality or area in a constituency.

The demand side of the problem is about the system not motivating or encouraging them to vote. There are politicians who suppress certain social groups from coming out to exercise their franchise. But the same politicians also motivate voters from their own communities to enrol and vote. Unfortunately, the Election Commission does not seem to have proactively stepped in to prevent such a manipulation of the electoral system.