Monthly Archives: April 2014

Imagining a City State

If you were given another life to live, it’s reasonable to pick James Lovelock’s.

Sometime ago he proposed that mankind should build more cities and hasten its migration to them so that we can get concentrated populations. His idea was that climate change is going to make saving all of the world impossible; so we might as well pick places we want to and do a better job.

Let’s assume one bought that theory for Tamil Nadu.

The state has a population of 72.14 Million. Its largest City, Madras, has a population of 8.91 Million in its metropolitan area. The metropolitan area occupies 426 square kilometers.

Now if we have citizens living in every other part of the state migrate to Madras and maintained its current density, the City will have to expand by about another 2400 square kilometers. That’s less than the size of the district that shares its borders with Madras: Thiruvallur.

Sounds like the most reasonable thing to do; that liberates all of the land in Tamil Nadu. We could also start the migration process district-wise to make it gradual. If a Chinese style order troubles our liberal conscience, there are always tax policies and the rest to incentivize instead of force. The incentives for public transport, movement away from feudalism and generally moving into the 21st Century all seem easier to design in this system. While it’s possibly also the biggest check on the anachronistic cultural hegemony that Brahmins in Madras enjoy.

And once every other state in India is reduced to a City, we can dissolve the nation state and have effective City states. Aren’t we all lapsed Romans, anyway?

Alliance Arithmetic in Tamil Nadu

Political alliances are a necessity in a multi-party FPTP system. But they come in a variety of shades. For instance the Left front can be treated as monolithic despite them having several constituent parties. But there are certain other alliances that are less cohesive and hence the voters’ stickiness or lack of it with a party as opposed to a coalition makes contests even more interesting. Tamil Nadu, with a largely bipolar polity and a host of smaller parties, offers one such experiment.

In 2004, it was a wave that washed out ADMK. The average margin of victory then in Tamil Nadu’s 39 constituencies was 162,533. So it’s not reasonable to consider it for our analyses which hopes to estimate relative strengths of parties. However, the 2009 Lok Sabha Elections in Tamil Nadu were very close to a real contest with an obvious skew towards the winner. So, let’s consider these results.

The table below considers all 39 contests and then provides a head to head tally for individual contests[1]. For instance, the first row with ADMK v DMK gives

  • The average margin of victory for all ADMK v DMK contests in the column named thus, regardless of who won
      • The next column gives the average margin of victory in ADMK victories against DMK
      • The final column gives the average margin of loss for ADMK’s losses against DMK
      • This repeats for other contests as we traverse and go down the rows

Average Margin of Victory

across all Contests

Average Margin of Victory

Average Margin of Loss













DMK V Left
















INC V Left












The average margin of victory for the ADMK against the DMK, as it can be expected, is 50.7% lower than for a DMK victory against the ADMK. But notice how the margin gap between wins and losses for the ADMK dramatically falls/reverses when it contested against the DMK’s junior allies instead of the DMK itself. In fact, the ADMK’s average margin of victory is higher than its average margin of defeat against the INC. There were a reasonable number of contests between the ADMK and the INC in 2009 and thus these averages are not a one-off contests.

The DMK, strengthening our hypothesis, did not lose even a single seat to a junior alliance partner from the opposition camp. Significantly, its average margin of victory against the then most important junior alliance partner of ADMK — the PMK — was 105,536. That’s 35.7% higher than its average margin of victory against the ADMK. This extends to the next most important junior ally – the MDMK — as well[2].

In short, whenever the larger Dravidian party was in a direct contest with a junior alliance partner of the opposing camp, it performed better compared to the average of that election. Now consider the cases where the junior alliance partners on either side contested against each other. There were only 5 such contests. In 4 such contests, the margins were much tighter for the winner and these 4 were split evenly between either camp[3].

That brings us to the simple question of: what happens when the DMK and ADMK don’t ally with these smaller parties? The argument often laid out in favour of such parties is that their votes are concentrated and hence they deliver their own constituencies. The PMK which is most often cited thus, as we just saw above, performed the worst. Its average margin of defeat in areas where its votes are supposedly concentrated were much higher than the election norm.

The entry of DMDK as an alliance parter, one could argue, alters things. The DMDK in 2009 Lok Sabha Elections and in the 2006 Assembly elections did not enter into an alliance. The party did not finish second even in one of the 39 contests discussed above. However, it’s estimated that this party too has pockets of strength. But in the State Assembly Elections of 2011, the DMDK did enter into an alliance with the AIADMK. Just as the PMK’s performance was worse than the norm in 2009 defeat, the DMDK’s 2011 performance was also worse than the norm; except in 2011, that was in a winning camp and thus being worse than the norm still meant a few victorious seats.

A reasonable hypothesis therefore is: when faced with a choice between a Dravidian Party and others, the Tamil citizen has a bias towards any one of the two major Dravidian parties even when it happens to not be the preferred one state-wide for that cycle. In such a scenario, in a week’s time, the Tamil Citizen is mostly going to have both major Dravidian parties on the ballot along with a third choice. If this history is any indicator, the NDA alliance has an unpleasant contest on its hands.


[1] – The original data was from the EC though I’ve calculated and tabulated it differently.

[2] – We can ignore both the VCK and the BJP in this list given they were both significant in just 1 contest each.

[3] – The fifth was a contest between VCK and PMK and the subtext of Dalit assertion against Vanniyars in a constituency where over 32% of the population belong to the Parayar community — a Dalit caste.

The need for hypocrisy

Pointing to an opponent’s hypocrisy isn’t a valid argument and it makes for dishonest debating. Very few, however, let the opportunity go. Especially when the debate happens in a public or a semi-public forum and the debaters are, at least in part, playing to the gallery.

Political speeches are about creating cleavages in terms of both positions and persona. Therefore, epistemic integrity from a politician can be ruled out. However, what matters is, what ought to be the yardstick in measuring general argumentation? If the merit of argumentation is judged by the merit of the proponent’s morality, human motives will be aligned to explain one’s status quo and not take the argumentation to its highest epistemic value. In other words, not only does the use of hypocrisy as an argument make for poor argumentation but it also defeats its purpose.

That brings us to certain ilk. An example of it will be: the right wing obsession about how a Liberal does not come to their defence even when liberalism is at stake. And that, is peddled as the reason why liberalism itself is somehow the villain[1]. That the failing of the individual is just that and not of the idea is the lowest hanging fruit. But when that’s not realised it becomes difficult take the argument to: isn’t the condition imposed on the said person — liberal or otherwise — limit that person’s ability to explore thought one’s own life is not in agreement with? And therefore, shouldn’t we as a society not point to hypocrisy of argumentation as an argument, if we wish to progress? And isn’t all progress in thought, therefore, achieved by nurturing the environment that encourages thought? Not even if but especially when it’s hypocritical?

[1] – Swapan Dasgupta’s Twitter feed is an example.

Is Federalism a Victim of Tragedy of Commons?

Politicians of a certain ilk have a certain campaign method. This ilk is made of those who are right of centre or those who imagine a middle class constituency or both. Their stock method is to talk in terms of projects or project ideas. Classic examples of this kind include Narendra Modi and J Jayalalitha: they go to a given constituency and discuss what they can bring as heads of government to that constituency. In other democracies, this’d be called pork barrel and the fiscal conservatives will be outraged at such promises. In India, rarely is any attempt made to even camouflage it.

The use of fiscal powers in the future to buy votes in the present has two major methods in Indian elections. However, these two are viewed very differently. If the campaign makes a promise of delivering welfare schemes, it’s considered fiscally profligate. But if it offers a project of some sort to locals — such as promising a bridge or a factory — it passes for development.

The two approaches apart from being rooted in different economic philosophies, have a crucial political dichotomy that is overlooked by those belonging to the former ilk. Typically everyone but the INC in India hold to federalism as a plank. Some do so as a matter of core principle — like the DMK and the SAD — and others simply because the INC has been in power most often in Delhi and they find federalism a political necessity to seek space for themselves. The BJP belongs more to the latter category than the former.

Whatever the motivations of a philosophical position, what matters to voters is an element of internal consistency. Let’s consider the AIADMK for instance — their manifesto does make significant points on the subject such as rejecting the Raghuram Rajan Committee Report and seeking central grants be shared equally under Article 275. The BJP manifesto on the other hand mentions federalism only once; its commitment to it is without any concrete plans to strengthen it. But the real issue is how does a commitment to federalism square with the constituency level pork barrel promises? What if those domains that the promises are being made under happen to be state subjects? Or, ought to be?

The flagship programs of UPA, such as NRHM, RTE, NREGA are all serious usurpations of state powers; especially in areas where state governments should be investing and concentrating the most. That’s of course expected from the INC which sees itself as a party of government in the Delhi; it doesn’t see itself as a champion of federalism. The INC however doesn’t promise as much constituency level pork barrel and instead makes a sweeping capture of states’ rights. But what do challengers whose base is in the states, but are seeking power at the centre, do? For example, Narendra Modi’s mention of cow slaughter as a Prime Ministerial candidate in makes it absurd for the BJP to call itself a champion of states’ rights. After all, slaughter of cattle is strictly a state subject and any attempt to make it a national subject because one disagrees with what the state has done insults the given state’s political dignity. But he does it anyway because it’s a cleavage between his party and the incumbents in the state he’s campaigning in. Is a larger welfarist usurpation any better or worse compared to these targeted attempts? Is there a difference? And shouldn’t state Chief Ministers, especially those belonging to regional parties, know better than to co-opt the same techniques?

The entire campaign rhetoric of many candidates contesting Lok Sabha elections revolve around local issues such as a bridge, a flyover, water scarcity, garbage disposal and similar things that one’d expect in the manifesto of a Mayor. There is of course the MPLADS which may be utilised for smaller projects. But for most cases an MP lacks the ability to deliver outcomes. When a leader of a Party, such as say Narendra Modi, promises such local outcomes it may carry better achievability. But the question, should he? What does the voter in a poor country do? Gobble up whatever may be the promise, given outcomes are rare, however distorted the source may be? Or, seek propriety in sources because that’s in the long term local interest? It appears there is a strange sort of tragedy of commons that forces less and not more federalism at least when it comes to campaigning.

Imagining a better Parliament

The individual citizen is largely powerless in densely populated Democracies like India. We’ve seen repeatedly —  across many methods of calculations — that the probability of impact of any one person’s vote on the outcome is so small that it’s practically 0. The actual outcome of that experiment, an MP, then goes to Parliament to vote on laws representing the interests of these individual constituents. However, the interests of these constituents are subsumed by Party interests in the actual representation. This is true in most Parliamentary Democracies but it is taken to the extreme in India by the existing anti-defection laws. Thereby making a citizen’s already minuscule impact far worse. Is there a better way to do this?

The obvious low hanging fruit is to do away with the anti-defection law. It was enacted to prevent legislators who jumped parties to bring down or form governments motivated by outright bribery. An MP switching parties for personal gain is clearly not in the interests of the electorate; though a less destabilising but far more frequently occurring and therefore more impactful aspect is: MPs belonging to a given party have to vote the way in which the party decides on every single piece of legislation even if it’s not in the interests of the MP’s constituents. On the other hand, if no form of control existed for their voting, it’d become a place analogous to the American Senate and it’s impossible to run a government elected under the Parliamentary system if the government is not sure if any of its programs will pass. Or, if the government itself will survive.

One approach that comes to mind is: allow each elected MP a fixed number of votes that can go against that MP’s party vote. That way if an MP feels strongly enough about something or it matters enough to that MP’s constituents but isn’t in the larger interests of the Party, a conscience vote gets cast. The problem with this approach, however, will be that it could result in the exact sort of scenario discussed above: say a majority of MPs belonging to the ruling party decide to exercise their “no” vote on the Budget; all at once. To guard against this, the requirement could be that such voters will have to register with the speaker of the house a day prior to the vote and only 10% of any party can do this for a given vote on a first-come-first-served basis. While this seems reasonable as a system of voting, the downside is that the person indulging in it is likely to hurt one’s own career and therefore professional politicians wanting to climb the party ladder are unlikely to do it.

A less palatable choice is that for every X votes an MP has participated in and voted along with the party, the MP gets n secret ballots allotted. If an MP after putting in the yards of several votes feels strongly about a given issue, that MP will have the choice to secretly cast one or all n ballots against the party’s line. This after casting his/her original vote along with the party. Maybe the speaker’s office has a secret booth akin to a confession chamber for this. That’d mean no vote is final until the speaker declares it is. Such a system is vulnerable to chances of corruption on the part of the speaker. The ratio of X:n becomes crucial as well and one can’t think of a ready ratio to suggest. More significantly, the electorate has no way of knowing who vetoed the legislation and that may become highly susceptible to corruption as well.

Almost all the choices above — from secret ballot to complete party-line voting as it now happens — have one problem or another associated with them. Worse, all options that veer away from party positions tend to some form of obfuscation from the electorate that is antithetical to the idea of representative democracy. After all posturing and perception are as important or even more so than the actual vote; that’s made impossible by secret ballots.

A solution one is tempted to explore is the structure of the Parliament itself.  A new electoral system where 1/3rd of the Lok Sabha is elected every two years for a six year term seems like an interesting option. This will achieve many likeable aspects without us even trying. Top among them will be the elimination of the obsession with the Prime Minister’s chair, given no election will be about that exclusively. But most importantly, it’ll likely solve the problem of representative voting discussed above because it’ll be interest of the parties to be more responsive. Voting constraints on the MP under this system will still be along party lines. But because there is a geographically equally spread election due every two years, a natural pressure of the representing constituent interests builds up to a slightly higher order on the party. That is, assuming the natural variation in “serious issues” tend to be normally distributed across geographies over time. Something along the lines of David Runciman’s hypothesis.

The other benefit of such a system will be that it’d make elections more a process and less an event; one hopes demagoguery of the sort one has experienced in 2014 will not have an incentive under such a system.

Swing Vote: Salem’s Vanniyars are its Germans

Let’s consider the Salem Lok Sabha Constituency in Tamil Nadu. The four major contestants are: V Panneerselvam from the AIADMK, S Umarani from the DMK, LK Sudheesh from the DMDK and Mohan Kumaramangalam from the INC. Of the four, two of them have family ties that largely explain their ticket to contest from their respective parties. LK Sudheesh is the brother-in-law of DMDK’s President, Vijayakanth. Mohan Kumaramangalam is the fourth generation politician from the Kumaramangalam dynasty.

Given the increasingly prominent role of Premalatha Vijayakanth in the DMDK at the expense and exit of the party’s founding mentor Panrutti Ramachandran, LK Sudheesh’s candidature is a strong one at least as far as the DMDK is concerned. He has also co-produced many of Vijayakanth’s movies in the last decade implying a degree of economic inter-dependence extending beyond family ties. So it’s reasonable to expect the party to spend a disproportionate degree of its resources in Salem.  LK Sudheesh has also come up with a manifesto of sorts for Salem. The the 33 point-plan reads more like that of someone contesting for the City’s Mayor than that of someone contesting to be a national law maker. It includes garbage disposal plans for Salem, railway lines running through Salem, flyovers, over-bridges, under-bridges and every other form of pork barrel promise that’s conceivable.

Mohan Kumaramangalam is a true local dynast who grew up elsewhere. His great grandfather, P Subbarayan, was a Zamindar and lawyer from nearby Thirechengode. Subbarayan was Madras Presidency’s Chief Minister in the 1920s. His grand father, also Mohan Kumaramangalam, was a Communist and later a member of the Congress Party and Indira Gandhi’s cabinet. His grand aunt, Parvathi krishnan, was a Communist leader and MP of some repute and enormous charm. His father, of more recent vintage and in living memory, was the former Power Minister who died in office. With all this though, what Mohan Kumaramangalam’s pitch to Salem is unclear. He has degrees from a couple of American Universities. His local pamphlets don’t read like a Mayor’s manifesto as in the case of his DMDK opponent. He is one quarter Gounder, the local land/industry owning caste, if that’s of any significance in Salem. So far, he seems earnest on the Internet and seems like one of us — a Macaulay’s child who wants to dabble in a few things.

AIADMK’s V Panneerselvam is an odd choice. The sitting MP from Salem, the veteran AIADMK leader S Semmalai has been denied a ticket. Semmalai, another Gounder leader, has a following of his own and he proved that to no less a person than MGR when the latter denied him a ticket in 1980; Semmalai contested as an independent and won. Of course this is 34 years later and whether that’d affect a rather nondescript V Panneerselvam in unknown at this point. Semmalai isn’t contesting this time and doesn’t seem to be implicitly supporting anyone either. The DMK has fielded S Umarani, one of the very few women contesting in Tamil Nadu. Both the candidates from the major Dravidian parties — V Pannerselvam and S Umarani — have almost no presence on the Internet. That must be reason enough to suggest these two are possibly the front runners.

While Kumaramangalam is a true political lightweight, DMDK’s LK Sudheesh isn’t. But the biggest problem Sudheesh faces is that the Salem constituency was originally allotted to the PMK in the NDA alliance but was taken away from them and handed to the DMDK. The two parties, DMDK and PMK, are natural opponents in the region and a voter for one party is highly unlikely to vote for the other; especially when there is bitterness of this sort. The internal alliance conflict of the NDA is possibly most visible in Salem. So, if the Vanniyar votebank of the PMK is a real thing, it’s likely to consolidate behind either of Dravidian parties. That may prove to be the differentiating factor. Whether the AIADMK knew this and therefore took a chance with Semmalai will make for interesting gossip.

So let’s assume between the DMDK, INC and independents, about 40% of the votes get spent. If that is so, the AIADMK and DMK have a contest within a universe of the remaining 60% vote share. Salem has 11.7 lakh voters and that makes the main contest restricted to just over 7 lakhs. It’s reasonable to assume anyone who gets about 3.5 lakh votes will win. If we therefore plug this N in our earlier calculation to be 3.5 lakhs, the probability of each voter impacting the election comes to e^(-292). However, if the Vanniyars vote as a block and are really the swing vote in the above scenario — their localised N maybe as low as 1 lakh — the impact of each Vanniyar vote goes up to e^(-157). That’s almost as high as that of a citizen in a constituency in Germany.