The futile quest for a perfect electoral roll | Number Theory

.

The Election Commission of India (ECI)’s Special Intensive Revision (SIR) of the electoral roll in Bihar has created a big political controversy. Opposition parties have expressed apprehensions that the strict documentation criteria which has been defined for being a part of the electoral roll can end up disenfranchising many people who do not have such documents. To be sure, the matter is now being heard by the Supreme Court and one will have to wait for a final decision. Without

Voter awareness e-rickshaws in Patna. (ANI)

The futile quest for a perfect electoral roll

• 
  Registered electors have generally been higher than the eligible number in the census

  HT has compared the number of registered electors in general elections which were held close to the census and compared them with the number of people eligible to vote in that election according to the census . The voting age in India was 21 years until the 1984 election and has been 18 since the 1989 elections. A comparison of five such pairs – 1971 election and census, 1980 election and 1981 census, 1991 election and census, 1999 election and 2001 census, and 2009 election and 2011 census – shows that the number of electors has generally been more than the eligible adult population, although the gap did come down in the 2009 election and 2011 census pair. These comparisons have been adjusted for states which did not have elections/census in certain rounds. The number of extra electors in these pairs ranges from 3.59 million (in the 2009-2011 pair) to 57.49 million (in the 1999-2001 pair).
• 
  States show a large variation in voting age population by census and elector counts

  If one were to take the latest census-election pair, namely the 2009 election and the 2011 census, states show a large variation between voting age population derived from the census and what is actually recorded in elector lists. For example, Haryana and Tamil Nadu had 20% and 15% of their voting age population by census missing from the electoral roll while Uttar Pradesh had around 12% more people on the rolls. In absolute terms, these numbers would be 7.2 million for Tamil Nadu and 12.2 million for Uttar Pradesh.
• 
  Mortality may cause the variance, but shows up even when census and elections were held in the same year

  What explains the large gaps between the voting age population as seen in the census and the electoral rolls? A reason for lower count on the rolls could be that some people do not register for elections. However, the only scientific explanation for more people on rolls than estimated from the census (at the national level) could be temporarily high age-specific mortality (since adult population has not decreased in between two census years in the data analysed here). One way to overcome the mortality problem is to compare electors and voting age population in states with assembly elections in the same year as the census. This should minimize gaps because the census has had a reference period of March 1 since 1991 while one needs to be an adult as on January 1 to be on the electoral roll. However, this does not show a better match between the two numbers, although electors on the roll were lower than that of adults in the census across the five state elections held in 2011.
• What does all this mean?

  The easiest thing would be to claim there is election fraud. The more sensible thing would be to accept some statistical discrepancy in exercises which involved hundreds of millions of people in the past and close to a billion people now. It is possible that census numbers are not hundred percent accurate, especially since they are self-reported and the census itself has flagged inaccuracies in single year age returns in the past. It is also possible that electors were registered at more than one place in the past but were enumerated in the census only once, which would show a higher elector count than the adult population in the census. Is this discrepancy worth putting the difficult burden of documentary proof on electors, or magnifying small discrepancies that brand the entire electoral process as compromised? This question is more political than legal in nature and requires a broad-minded approach.