Monthly Archives: July 2013

Can the BJP game Tamil Nadu?

AB Vajpayee lost in 2004 mostly because his party made a poor choice in Tamil Nadu. And not because India Shining is easy to blame. The BJP broke its alliance with the DMK(then a member of NDA) and instead allied with the AIADMK[0]. And given the decisive nature of Tamil Nadu’s mandate over the past 2 decades, that meant forfeiting 40(39+1) seats in what was essentially a hung Parliament.

Since that disastrous result for both AIADMK and the BJP, both parties haven’t allied with each other in any of the elections so far. Further, no major party has allied with the BJP in TN. While the INC is not a major player in the state either, it has managed the 40 seats better than the BJP in the past decade. Had it not been for decisive Tamil votes, both UPA terms wouldn’t have been possible.

It is reasonable to expect that the UPA will not be a beneficiary of a third successive vote from TN both from a historical perspective and from that of recent opinion polls. That being the case, what does a BJP analyst do? Facts to consider are,

  • Tamil Nadu’s Muslims form about 5.56% its population. That is quite a small proportion of the population compared to most large states. West Bengal, Uttar Pradesh and Bihar for instance, where the Muslim vote is a significant political phenomenon, have 25.25%, 18.5% and 16.53% of their respective populations constituted by Muslims.
  • Tamil Nadu is probably experiencing the nation-wide upward trend in the BJP vote in this election cycle, according to the same opinion polls. This when placed over and above a low base now possibly puts the party, after all considerations, at close to 10% vote share. That’s an absurdly high percentage for the BJP in TN to squander by fighting the elections alone. Should it?

The BJP is obviously closer to AIADMK than it is to the DMK both in terms of personalities and ideology. So as a party that has an improving vote share in the state, the BJP should like to position itself as a conventional electoral ally to the AIADMK. Not only does this have the benefit of improving the probability that the NDA would be called first by Pranab Mukherjee but also in terms of legitimizing the NDA as a viable alternative to voters and alliance partners in other states possibly wondering about viability of NDA even if the UPA were a lost cause. However, in doing so, the BJP enthusiast, and more specifically that NDA ethusiast, is missing a vital aspect of this game. The NDA is unlikely to form a government without the AIADMK. Therefore it’s in the interest of the BJP and the NDA to maximize the AIADMK tally, especially because they come in seats the BJP is unlikely to win on its own anyway. However, if it signals explicit cooperation in any way, that will only have the opposite effect on the actual seats that the AIADMK wins. This likely to happe for two reasons.

Firstly, the Tamil Muslims who’ve historically voted DMK in large numbers have slowly been moving towards the AIADMK for various reasons[1]; needless to say this blunts a crucial DMK advantage that no long-term AIADMK strategist will want to let go. The two major Dravidian parties have had such stable vote shares for so long that any movement in a segment, such as Muslims, is likely to have serious long term impact. A post poll alliance that weighs the long term prospect of an election few years away against immediate power possibly looks more kindly at ‘vote-bank’ losses[2]; for every politician has to believe there will be an overall upward move in the next cycle immediately after an election to remain in the politics business. But a pre-poll understanding has no such luxury of long-term optimism and therefore is a much more difficult bridge to cross. In this case one that surrenders an accruing advantage for no real trade-off. The BJP’s upward trend is most likely a one-off event in TN as it stands. To surrender a long term advantage to a questionable flash in the pan, especially in an election one might anyway win, is a poor choice to make for the AIADMK.

Secondly, of all the local partners available in Tamil Nadu the Left parties that are still technically an ally of the AIADMK have the best ‘transfer-ability rate’. That is, an alliance with the Left uniformly transfers their votes to their alliance partner across the State at the highest rate compared to every other junior alliance partner in Tamil Nadu. This has been true in the past three election cycles. It’s also a simple truism of Indian politics that the BJP and the Left aren’t going to be in the same team. So, by allying with the BJP, the AIADMK would lose any prospect of a tie-up with the Left. Worse, pushing it to the DMK camp. Pushing a possibly existing alliance partner with a known high transfer-ability to the opposition camp for a partner with an unknown transfer-ability rate will be a gamble unbecoming of a serious political party. Worse yet, the pre-poll alliance with the Left or a completely independent AIADMK leaves open a low probability event of Ms Jayalalitha becoming a Prime Ministerial candidate after the elections. That’s almost reduced to 0 with a pre-poll NDA alliance.

So, if what’s in the AIADMK’s interest is also in the BJP’s interest, the latter would simply shut up and let events in Tamil Nadu flow. If this were a one-person game, that’d be far easier to do. But BJP’s political opponents at the national level will raise the issue of ‘where will you get allies from?’ knowing quite well the answer will ruin the intended result. It’s a far more difficult wager to swallow when your silence may or may not yield results but your lack of it will only dent your own chances. Does the BJP have enough discipline?

[0] – It’s debatable who is to blame for the breakdown. Let’s assume the BJP did it for the purpose of this post as the DMK accused it of hobnobbing with the AIADMK at that time.

[1] — That’s an assertion with a lot of recent political history which deserves multiple posts of its own. But the nature of protests in Muslim causes in TN are a symptom of this phenomenon.

[2] — It’s not certain if that’s true in Andhra Pradesh where one could fit Chandrababu Naidu in the category if it were true. A detailed look at data on that count is warranted.

Methods & Outcomes

Modern academia doesn’t have place for a polymath. But in the 19th Century, most Scientists were that for the obvious reason that most modern academic disciplines did not even exist as a separate branch of study. And the best illustration of this epistemology is the story of Alfred Russel Wallace. Yes, that forgotten contemporary of Charles Darwin from whom Darwin most likely plagiarised some elements of his work. But more importantly for this post, towards the end of his life Wallace joined the anti-Vaccination campaigners and published in the Lancet suggesting small pox vaccinations did not work if one were to consider the outcomes.[1]

Evidence based medicine which we now take for granted[2] was sort invented in this period and Wallace was possibly one of first few pioneers to bring a certain rigour to this discipline. Of course he was wrong as we now know but in his environment, given the data he had, he was correct; in that, in the poor sanitary conditions and its equalising effect on 19th Century England, vaccinations had no clear positive impact. And the reason Wallace is a truly great Scientist is because his method was right though his conclusions were disputable at best and mostly wrong in retrospect.

The other point that Wallace’s Lancet publication raises in retrospect is disprovability. All true sciences have that for a foundation while the soft sciences still allow not merely for the absence of a control group but sometimes even selective use of Data; such as in Economics. To cite an oft quoted example, two Economists who argue about trickle down theories can point to USA and China over the past 50 years and arrive at opposite conclusions. That still does not invalidate the discipline just as it brings into focus the even greater search for other contributing variables which may explain this phenomenon better. The moral of the story here, just as in the case of Wallace, seems to be: the correctness of the actual conclusion at any point is immaterial in the long run as long as the rationale of the method has a sound scientific basis. In other words, the disputable conclusions on Gujarat and Kerala are not reasons to run down one of the last living genuine Polymaths, even if they end up being correct.

What’s a far easier conclusion for a lay person to make however is that of a moral imperative. One of the sacred covenants between the ruler and the ruled in a modern Democracy is that an extension of State’s patronage is not an expectation of pliancy. That was broken by another scholar whose work seems serious enough to have warranted an understanding of rational method.

[1] — If the source where Alfred Wallace’s work is linked to worries you, it worries others too.

[2] — One hopes modern medicine is evidence based after all that. 

Converting Votes to Seats

The charts below illustrate the the relative seats compared to vote share for the top National Parties in the 2009 General Elections. The second chart below looks at the hit rate of the parties in terms of how they convert votes into seats. That is, it shows how many seats each party has for every percentage point of support it has won.

Vote Share vs Seats Won -- 2009 General Elections

Vote Share vs Seats Won — 2009 General Elections

Hit Rate for converting votes to seats, 2009 General Elections

Hit Rate for converting votes to seats, 2009 General Elections


Are regional parties a bad thing?

Elections in India, as almost everyone is aware by now, follows a ‘first past the post’ system. It’s a deeply flawed idea that often throws up results that do not represent the true will of the people as elections are supposed to. The most common quirk of this system is that in a large and plural democracy like India it encourages the tyranny of an even smaller majority (caste) than the one faced by other small western nations that have adopted some form of Parliamentary democracy. This post will attempt to examine how it works as a barometer of “true will.”

To begin, let’s point to the greatest flaw of this system: the post in the ‘first past the post’ is not really a post. That is, it’s arbitrary and decided after the event. It’s not set at say 50% which contestants have to attain in order to win. In a tight four cornered race, usually the winner has less than 30% of the votes. In effect, a political party is left with a zone of vote-share optimality in order to maximize its seats won. Let’s take an example to prove the point. Consider a constituency like Vidisha where Ms Sushma Swaraj secured 78.8% of the votes while the closest rival had 8% votes. This means a lot of voters who voted for Ms Swaraj just wasted their support for BJP in one constituency and the party would have been better off had a chunk of them been spread out uniformly across the country. On the contrary, if every BJP vote had been so uniformly spread that in every constituency it contested, it had its vote share equal to that of its national average at 18.8%, it would probably have lost almost all its seats.That leaves the BJP wanting to operate in a zone where they do not want additional support in paces where it isn’t needed any more and yet don’t spread it too thin that they aren’t winning the marginal contests.

If we take the 2009 election outcome and compare it with the vote share, what emerges is that the two are correlated reasonably fairly. The INC won 28.8% of the votes and 206 seats — that’s 38% of the seats were won with 28% of the votes. Which is not such a bad distortion for the leading party. The BJP won 21% of the seats with 18.8% of the votes. As one goes down this path, the ‘hit rate’ dramatically reduces for smaller parties. Does that make Lok Sabha a much fairer place compared to say the House of Commons where the distortion is far worse? Let’s look at that in some detail in the next post.