Monday, May 30, 2011

Did TN swing away from the DMK or not?

There was an article in here, a few days back. Titled "Tamizh voter did not punish the DMK. Here is why", it was written by Mr.N.Gopalaswami (a former Election Commissioner of Tamil Nadu) and Praveen Chakravarty.
The article got a lot of comments and I realized its popularity when a friend mentioned it to me in a meeting. I had myself written a comment to that article, buried somewhere - I had argued that the authors were wrong.

This is an important talking point - and we have to be careful about it. What the authors purport to prove is that in spite of the coverage of the corruption cases against A.Raja and evidence of DMK involvement in the 2G scam, the people who vote for DMK continue to do so. But from that, they jump to the conclusion that the TN electorate has not punished the DMK. In other words, corruption is not a big issue in people's minds. Implicitly, being corrupt does not seem to be punished by the electorate.
This is a very serious result, if true.
I decided to do check the numbers myself and analyze the results. This post is a description of that.

The Authors' argument

In 2006 TN assembly elections, the authors say, of every 100 people in a constituency that had a DMK candidate, 46 voted for the DMK.
In 2011, 42 voted for the DMK.
Therefore only 3 people swung away from the DMK; and this is a very small percentage. The authors argue that the election results wound up the way they did, with a massive ADMK win, only because of the alliance between DMDK and ADMK.

Adjusted Vote Share vs Total Vote Share

Here are the links to the wikipedia pages on 2006 TN assembly elections and 2011 assembly elections.

You will see in the bottom of the page, a table with the final breakup of votes. You will see a column called "Vote %" and another called "Adj Vote %".
The DMK's vote percent in 2006 was 26.5%. In other words, out of 100 people who voted in that elections, only 26 voted for DMK.
The DMK's vote percent in 2011 was 22.4%. Out of 100 people, 22 voted for DMK.
This means that 4 voters swung away from DMK in 2011. But what is more important, this 4 as a percentage of DMK voters is 20%.
So 20% of former DMK voters did not vote again for DMK. That is a pretty significant percentage.

But where did the authors of that article come up with the numbers 46 and 42?
They used the NEXT column called "Adj vote %". They used the adjusted vote share, which is the average of the vote shares per constituency.
I think the total vote share is more important than the adjusted vote share. By the total vote share 20% of DMK voters swung away. By Adjusted, less than 10% swung away. I think this is a pretty significant difference.

But let us assume we use adjusted vote shares. All that it means is that the voters of DMK continued to vote for it. That does NOT mean (as the authors imply) that Tamil Nadu electorate had not punished the DMK - people who did not vote the last time, and new voters may have opted for the opposite party, as I will show below.

Swing in favor of ADMK

In 2006, the total vote share of the ADMK was 32.6%.
In 2011, it was 38.4%.
In other words, 6 voters out of 100 swung towards ADMK. The authors have simply not counted this as a reaction against DMK. But, in an almost two party system like Tamil Nadu, the swing in favor matters.

Absolute Numbers
Most importantly, the voting percentage in 2006 was 70%. In 2011 it was 78%. This jump means that the total number of voters who voted for ADMK went up from 10.7 lakhs to 14.1 lakhs. While those who voted for DMK went down from 8.7 lakhs to 8.2 lakhs. Part of this has to do with lesser number of constituencies the DMK competed in (130 in 2006; 119 in 2011). But the difference is just 11 seats. That cannot account for their voters number not even going up with the increased electorate's size (Total votes were 32.9 lakhs in 2006;36.7 lakhs in 2011).

In conclusion, I would say the DMK WAS punished by the Tamil Nadu electorate. I think people are smart enough to realize when they are being conned.

No comments: