A Flawed Democracy – The Case for Proportional Representation in India: Srinivasan Ramani


Times of India graphic

For all the chest thumping and tomtoming about the Samajwadi Party’s emphatic victory – winning 224 seats out of 403 in the UP Assembly elections – a true reflection of the mandate is to be seen in the individual vote shares of the four main (“effective”) parties in the elections (the Bahujan Samaj Party – Bhartiya Janata Party, the Samajwadi Party, and the Congress – in alliance with the Rashtriya Lok Dal). Data from the Election Commission of India website shows the following in terms of vote shares:

In other words, a mere vote share of 29.3% was enough to garner 226 seats for the SP (see figure 1 for the discrepancy) while the BSP was to be satisfied with only 80 seats despite garnering only about 3.4% lesser than the SP. See this chart:

To be fair to the SP, the 2007 assembly elections also had a similar picture as shown below:

The 2012 Assembly Elections

A closer look at the election results – which was a tight four cornered contest in many constituencies – is warranted, because of the low swing in the favour of the SP that still enabled it to win very handsomely seat wise. After a detailed look at the numbers, we shall then proceed to substantiate some normative arguments about the party system, the electoral system and how this matters in terms of the political economy of the state.

Candidates secured more than 50% of the overall votes in only 16 constituencies. The majority of the victors had less than 40% of the overall vote; in fact close to 117 winners polled less than 30% of the overall vote. This plot provides a decent picture of the range of winner’s polling percentages across constituencies:

The fact that this was a very close contest in most constituencies is evidenced by two sets of statistics – the ratio of the votes won by the loser to that of the winner (Loser-Winner Ratio or LW ratio) and the ratio of the votes won by the second loser to that of the winner (TW Ratio). If the LW Ratio is close to 1, it means that the margin between the winner and loser is very close. And if the TW Ratio is close to 1, it means that the contest is a very tight three way race.

A plot of the range of LW Ratios versus the % number of constituencies having such ratios gives us a good picture of the closeness of the contest –

Note that, in close to 60% of the constituencies, the (first) losing candidate polled more than 80% of the votes tallied by the winner.

Note here that the in close to 10% of the constituencies (about 100 seats), the third placed contestant polled more than 70% of the votes won by the winner.

In whatever way we slice and dice the data, it is pretty much clear that the contest was pretty much a four cornered fight between the BSP, the SP, the BJP and the Congress-Rashtriya Lok Dal alliance in most constituencies. Despite having close to 225 parties (and many other independents) in the fray in the constituencies in the state, about 234 seats out of the 403 saw a direct fight between the four parties for the first four positions.

However, as this site points out, what marks the SP as better than the BSP is that it has defeated the BSP more comprehensively in many constituencies as compared to the vice versa. To quote the analysis from “The Open Data” website –

If one considers a simplified electoral contest in all the ACs [Assembly Constituencies], reduces them to face-offs between 2 parties (only winner and runner up) and analyzes the data for the Samajwadi Party, one finds that SP and the BSP fought head-to-head in half of the assembly constituencies, with SP winning majority of the time (SP’s strike rate against the BSP was 79% with a 8% margin of victory); BSP won only 43 out of 204 ACs against the SP with a 4% margin. When facing off against the BJP, SP’s strike rate was lower at 65%, but the median margin of victory was higher at 10%. On the other hand, against the Congress, SP had a lower strike rate (58%) and a lower margin of victory (4%).

Yet the unmistakeable conclusion is that despite the above, the rather close nature of the contest suggests that the eventual seat share of the SP is far higher than what was due to it if vote share was to be the basis of seat allocation. Without doubt, the First Past the Post system disproportionately awarded the SP, as was the case in 2007 when the BSP enjoyed that privilege. One needs to look at trends in UP elections over time to see if these two elections were aberrations.

Vote Share vs. Seat Share in UP

Has there been an increase in the tendency to reward seats disproportionately to parties with highest vote shares in the state? Here is a comparison of the vote share – % of votes won vs % of seats won in elections since 1989 –

Note that the 50% mark for seats has been breached only thrice in the seven elections held so far in the state since 1989. But that mark has been achieved in the last two elections with comparatively low vote shares – a trend that was visible in the 1970s in Uttar Pradesh.

But there are major differences between the 1970s and the elections since the 1990s. The ‘70s featured a time when the Congress was the largest political party in the country and it took the Emergency and the might of a combined opposition to overcome it in 1977, even though the vote share had jumped only about 5.7% from 1974 for the Janata Party combine. But again this election was virtually a referendum for/against the Congress and the single issue helped the opposition win a brute majority. Just as the experiment failed in the centre, the Janata combine could not manage to hold on as the Congress came back to power helped by a vote swing back in its favour.

The 1990s marked a pivotal change in UP politics with the emergence of the BSP and the SP inaugurating what is known as the “Mandalisation” era. These two parties effectively articulated strong Dalit/OBC-Muslim based appeals (respectively) to relegate the other “national” parties – defined here as having electoral relevance and reach beyond Uttar Pradesh – the Congress and the BJP to lower positions in 2000s. This process was not very smooth. Contiguous to the Mandalisation process was the effect of “Mandir politics” – articulated by the BJP and counteracted mainly by the SP and the effect of the third M of Indian Politics in the 1990s[iv] – the Market based liberalization reforms – which came to alter the political economy of the state drastically. The last of the processes saw the parties trying to articulate newer positions on “economic development” and governance. As Vivek Prahladan’s (2012) commentary mentions, the late 2000s have been a period where the “regional” parties are still rooted in their identarianism but have tried to posture about the need for “economic development” while the national parties have tried to adopt identarianism through their own biases in order to stay relevant in the state.

These trends have resulted in a political contestation in the state that is not limited to a narrow correlation between the strength of political forces and respective identity bases – castes/ communities/ religions – but a more complex web of support bases. Therefore the vagaries of the first past the post system has benefited the party with the marginally highest vote share, despite the relative effectiveness of support for all four parties in the state.

In fact contrary to the trends in UP, nationally speaking in parliamentary elections, the difference between vote share and seat shares has been reducing over time, as the country’s polity has slowly but steadily transited from a one-party dominant system to a raucous multi-party system that has made coalitions an imperative in the center.

Note: Except for 1977, 1989 (Bharatiya Lok Dal, Janata Dal), 1996-1999 (BJP), the Congress was the single largest party in all elections.

The accentuation of the vote-share/seat share divide is pretty unique to Uttar Pradesh, in the sense that this has occurred despite the lack of bipolarity which has generally been a characteristic of state elections in most states (and certainly larger ones in the country).

With the presence of the Congress (or its ally, the RLD) and the BJP (apart from the SP and BSP) as effective parties in most constituencies, it has meant that the SP could easily overcome the BSP in many constituencies despite a minimal swing in its favour. That both the SP and the BSP have distinct social bases in the state makes this character of the UP elections even more important to analyse.

Although detailed analyses of vote shares are not yet available (they should be available after CSDS’ Lokniti publishes its comprehensive post poll surveys which generally provide a compelling picture of the elections), it can be assumed that the Jatav community and other dalits once again voted overwhelmingly for the BSP while the SP retained its vote share among the Other Backward Classes and gained support from the Muslims.

Roughly, the BSP continues to articulate the interests of the Dalits primarily, but has enjoyed the accrual of support from some sections amongst the OBCs, Muslims and even the upper castes – a coalition that was certainly more numerous in the 2007 elections than what is expected to be in the 2012 assembly elections. The SP had retained its core support among the Yadavs and other OBCs as well as the Muslims (who still support the party due to its past positions and actions on the Mandir issue), but the swing this time is also a result of shift of “tactical” support from hitherto antagonistic sections among the upper castes. The Congress and the BJP have had bases of popularity among the upper castes, with the BJP enjoying some in regional bases (Bundelkhand?) in the state, with the former probably seeing a slight accrual of support from the youth cutting across various identities as well.

Studies (Chibber and Nooruddin 2004) have shown that states in India with a large number of effective parties tend to see selective and not universal delivery of public services, as the ruling party tends to service only select social bases.

One could argue that this form of a disproportionate allocation of seats relative to vote percentage in a highly contested polity is perhaps a feature only of UP as most other states now see two cornered fights between alliances. Tamil Nadu, another large state with a larger number of effective parties with large social bases of their own, invariably sees a two-cornered fight between fronts led by the two largest Dravidian parties. But these two Dravidian parties, in order to stay relevant electorally, have had to stitch coalitions with various political forces that represent different identity groups and political constituencies, thereby effectively ensuring that the polity is bipolar between two fronts led by the respective parties. This is unlike what is the situation in UP.

Case for Proportional Representation?

There would have been a truly representative verdict that allocated seats proportional to vote shares if the voting system was based on some form of proportional representation (PR) – mixed PR, allowing for transferable votes, etc. It has been argued that the FPTP (First Past the Post) system has run its course in India and that it is time that the country adopted PR systems that would allow for representation of minorities and smaller parties in the legislatures. However, proponents of the FPTP system argue that PR is a recipe for instability as exemplified by the current political deadlock in Nepal, a similar socially stratified country that has adopted the PR system. It is also argued that the FPTP system has not discouraged the growth of smaller parties as seen in the gradual regionalisation and federalisation of India’s polity. And that affirmative action in the form of reservation of seats for marginalised groups such as the scheduled castes and tribes as also the need to obtain support from diverse sections of the population has ensured desirable outcomes in terms of representation, without sacrificing too much on inherent stability as compared to PR systems.

But as UP shows, there are glaring cases of skewed verdicts in the FPTP system. UP has the largest state assembly and it sends the largest number of representatives to Parliament. If the electoral trends in 2007 and 2012 are repeated regularly and the larger parties remain oriented in their social outlook to a limited set of castes and communities, the clamour for a shift to the PR system will only grow.

(Srinivasan Ramani is Senior Assistant Editor, Economic and Political Weekly. Many parts of the article are also from a recent EPW editorial on the very subject. All tables and graphs in this article were created from data extracted from the Election Commission Website. For those interested in how this was done: the procedure was to download all constituency information (voting data) through the excellent Firefox addon – “Downloadthemall” and after writing a small batch script to extract the tables and make it readable in spreadsheets. There are other elegant methods possible – through python tools such as “beautifulsoup” to do so as well.)


Chhibber, Pradeep and Nooruddin, Irfan (2004), “DO PARTY SYSTEMS COUNT? The Number of Parties and Government Performance in the Indian States”, Comparative Political Studies, Vol. 37 No. 2

Prahladan, Vivek (2012), “UP Assembly Elections: Politics of ‘Belonging’ or ‘Belongings’?”, Economic and Political Weekly, February, Vol XLVII No.6

Yadav, Yogendra (1999), “Electoral Politics in the Time of Change: India’s Third Electoral System, 1989-99”, Economic and Political Weekly, Vol. 34, No. 34/35

From Kafila archives:

See also:

16 thoughts on “A Flawed Democracy – The Case for Proportional Representation in India: Srinivasan Ramani”

  1. Brilliant analysis..The winner-takes all system is really unfair.In the current system all votes for the losing candidates are ignored. This means that parties with only a slight majority of votes can end up with a vast majority of seats. Smaller parties may never get enough votes in one place to win a seat but may get hundreds of thousands of votes across the country, yet they are still ignored.It’s time to climb up to the PR system.

  2. Hi,

    Appreciate your analysis as it is much better to with support of data rather than just try to fit own biases into the result which is what media lead by TV channels does in India.

    In terms of solution for electoral reform however I would suggest that beside FPTP and PR there is another system which is a middle way and is fairer than FPTP but does not result in stasis like in PR.
    In this modified FPTP candidate is declared winner only if she is a winner by majority that is more than 50% of polled vote (only 16 in UP!) but the second round is held where only the winner and runner-up participate and winner is the representative. If this was case in UP then it would have been imperative for small player like BJP and Cong (in context of UP) to declare their preferences for their supporters which could have altered the election result.
    This helps specifically in multi-cornered election which produced hung verdicts as this would force people to chose their second best parties and would tend to reward the ones the moderates, extremes tend to do well in multi-cornered contests like BJP at the height of Mandir issue would have been untouchable.

    I believe Australia has a another version of the same where you have to rank candidates in the order of your preference which would be slightly cumbersome and complex process in Indian context.

  3. Impressive, unbiased analysis. I am troubled with the phenomenon of dynastic political succession. While Nehru-Gandhi dynasty has ruled India for quite some time, Nehru might have promoted Indira, but he did not appoint her a Cabinet Minister in his life-time. Also, Rajiv Gandhi came to power only after the assasination of his mother. On the other hand, JP’s successor Lalu installed his illiterate wife Rabri as Bihar CM, and now Lohia’s disciple Mulayam crowns his son as UP CM. Fortunately, Akhilesh has college degree. Hope, some sensible posting on this topic in Kaflia.

  4. There would have been a truly representative verdict that allocated seats proportional to vote shares if the voting system was based on some form of proportional representation (PR) – mixed PR, allowing for transferable votes, etc.

    When you have a divided polity, all systems will be flawed. Take a look at Israel which has such a divided polity that in desperation, it has tried almost everything. It still has a notoriously unstable political system.

    If we do move in the direction of PR, we won’t end up like Germany or any of the Western European democracies. More likely, we will end up looking like Lebanon which has a very intricate PR which even specifies how the top political posts are to be rotated among the three main religious communities. Do we really want to go down that route?

    The one advantage of the first-past-the-post is that it is simple both to understand and to administer. As it is, the administration problem is so severe that our Election Commission conducted the elections in UP — just one state — over a period of a month! Can you imagine the nightmare that will be involved in a move to something like a single transferable vote? To start with, how are the electronic voting machines going to be modified? And that’s just the start.

    There is one more thing about the first-past-the-post system. The fragmentation of votes forces political parties to reach out to other communities and to build broader coaltions. It may not have happened yet in UP (but then it did happen to an extent in 2007) but I think it will happen at some time. Look for instance at Punjab. Isn’t it striking that the re-election of the Akali Dal is partly to do with the fact that it has ceased being a party of only the Sikhs and has reached out to the Hindu community? In a country like ours, the last thing we want is for the polity to become even more divided.

    When we are a richer and (hopefully) more mature country, we can try things like PR. For the moment, let us leave things as they are. No, it is not perfect but it has worked — sort of, for sixty-odd years. Let us leave it at that.

  5. I’ve had it with the current electoral systems I would like to see a complete rewrite of how elections are managed and votes are counted. I’d like to see a new “Open Source Electoral System.

    An electoral system built by the people, for the people, to govern the people, FOR THE PEOPLE!

    I make my case and present my suggestions on a blog at http://onerealkewlguy.wordpress.com/2012/03/14/citizens-need-to-deal-with-electronic-voting-now-to-prevent-fraud/

  6. Something which I’ve been forwarding to a lot of people following the Punjab Polls:

    I think it highlights an inherent flaw in the first-past-the-post system of electoral representation. Our citizenry is pliant, self-serving and parochial. This produces a fundamental bias of representation in our democracy. Haven’t the expectations and awareness of the venal electorate been overburdened with a cultural affinity towards tribalism and a misplaced sense of history? I’d written a small piece to a political leader in the same regard, inspired by the referendum held in the UK on choosing the proportional system of representation, which was rejected by its peoples. Please bear with my indulgence:


    “A heartfelt note written to a friend, who is also a senior figure in the Punjabi political establishment. UK had a referendum last year to choose the proportional system of electoral representation instead of first-past-the-post, which was rejected by its peoples. Is the rise of caste politics and civil disobedience movements like that of Hazare’s pointing to an inherent bias of representation in our democracy?”

    Dear Mr. XXX – As you may better understand, such prescience and lucidity of thought does not percolate to the higher echelons of politics. Not that the leaders are incapable of having the much-vaunted depth and foresight (as is generally assumed), but these existential observations necessitate one to be a mere observer, dissecting the situation with the scalpels of intellect. An empathic doer, on the other hand, would rather choose to succumb to the expectations of his people. It is the same susceptibility and an innate ability to tap the collective consciousness which rewarded him with the mantle of leadership in the first place. As it often happens, this overarching responsibility and the proverbial bias of representation which comes with it can overshadow moral righteousness and prudence. This peculiar dichotomy can best be seen as the trappings of populism. However, there comes a time when the flow of rightful democratic expectations ebbs, when the order of enfranchised representation should be reverted, when a balance has to be struck between the rights of an individual and his social contract with the state, when ideological course correction should take precedence over populist appeasement. For it is not only the leadership which falters, the electorate can also fail to understand the pluralistic aspirations of the state or overburden themselves with a misdirected sense of history. Conscious of this stark dichotomy, the leader should take the brave and selfless decision of calling for an overhaul, thus rising to the hallowed pedestal of a statesman rather than a mere politician. The zeitgeist in Punjab is calling for such an inward and inclusive overhaul.



    Before we start howling about the “marauding” Akalis thrusting their way to power, our ignominious existence as the citizenry that is pliant, self-serving and parochial must be laid bare. It doesn’t need a master psephologist or other such creation of a hopelessly-objective world to realize that the incumbents bludgeoned the opposition with a masterful blow of money and hegemony.

    Any attempt of blaming or loud wailing must begin by regretting our indulgences, like counting the number of mistresses that ruined one’s family life during that moment of clarity. The most pervert attachment that becomes immediately noticeable is with the media — another cabal that leverages the power of the institution to turn truth into propaganda. How many regional newspapers have become the bitches of their political masters? Ajit, Tribune, Jagbani – you name it. How long will this attempt of self-censorship hide the smell of stained and rotting newsprint that is embellished and perfumed with the illusions of truth, labeled innocently as paid news? Sukhbir’s road trips with the clueless national media were remarkably candid and revealing when he hinted that the realities at the grassroots are completely different (for good or bad) from what is being portrayed, that it is only a matter of time.

    As one adjusts to this surge of conscience, it would be apt to highlight the fact that we cannot put the blame solely on the machinations of the ruling party. Haven’t the expectations and awareness of the venal electorate been overburdened with a cultural affinity towards tribalism and a misplaced sense of history? While interacting with a senior Akali leader last year, when their chances of winning looked dim, we shared a common conclusion that Punjab also needs a good and diligent opposition; a shadow cabinet that would keep the government on tenterhooks. Development cannot and should not be directly equated to rule; projects and ideas can also be nurtured by pushing forward an incisive debate that the establishment cannot escape from. But then the question arises, is there any one left “clean” enough to take that charge? All the closets are full of skeletons.

  7. I feel that this article attacks the issue of representativeness of Indian democracy in the crudest terms: i.e. what is the relationship between seat-share and vote-share? But a government that is cobbled together with a lot of regional parties is not necessarily a representative government. How else do you explain the ability of a succession of minority or coalition governments to push through neo-liberal reforms?

    1. But a government that is cobbled together with a lot of regional parties is not necessarily a representative government.

      Okay, I’ll bite: why not? Because they implemented policies that you don’t agree with? In that case, can I say that the entire Jawaharlal Nehru and Indira Gandhi eras were totally unrepresentative?

      You cannot define “representativeness,” by looking at outcomes, whether it is the ability to implement neo-liberal reforms (whatever that means) or anything else. By the way, how do you know that the majority don’t want “neo-liberal” reforms?

      “Representativeness” is a very problematic concept, especially in a divided polity like ours. Social choice theorists generally proceed by postulating some “reasonable” axioms that a “representative system” ought to satisfy and identifying the voting system that is consistent with their chosen set of axioms. But by varying the axioms, one can get different systems. There is nothing to say why one set of axioms is “better” than another equally appealing set of axioms. And unfortunately, there is no voting system that satisfies all the desirable properties that we might require it to satisfy.

      Pratap Bhanu Mehta puts it really well in this clip: “The odditity about justice in representation is that there is no representative process that can decide what is justice in representation. On the face of it, one can have four normative theories, equally appealing…” (See remarks towards the end of the clip, as indeed the rest of his very thoughtful remarks.)

      1. “Okay, I’ll bite: why not? Because they implemented policies that you don’t agree with?”

        You’re right. I have no special knowledge of people’s preferences. The pursuit of “reforms” despite mass protests gives me the impression that governments are not necessarily “representative”. For instance, it is hard to imagine that FDI in retail has “popular support”. I don’t know for sure, of course.

        Your latter point is well-taken, but I think it supports my point that all forms of government are imperfect forms of representation of the “people’s will” (assuming such a thing exists). For instance, to use Amartya Sen’s example – assuming there is a general preference for ensuring that no one dies of hunger-related illnesses, authoritarian regimes are better at minimizing routine hunger-related deaths whereas more democratic regimes are better at preventing famine conditions from excalating. Merely ensuring that there is no mismatch between seat-share and vote-share does not automatically result in better translation of “public choice” into “public policy”.

  8. Very interesting article and a fine analysis by Srinivasan Ramani.

    An aspect to be kept in mind is whether the voting pattern will remain the same if we move from First Past the Post to Proportional Representation. The present article assumes that the voter will continue voting the same way. But it is possible that voting behaviour will change if we move to PR, i.e. voters will play the system differently. It will be interesting to study the voting patterns in other countries which had the FPTP earlier and then PR to see if this took place.


  9. the PR system will also provide an opportunity to the fringe mass environmentalist movement in this country to make a mark in electoral politics. At the moment the mass bases of organisations like National Alliance of People’s Movements are so small that they cannot win any seats apart from at the panchayat level. However, if there is proportional representation then these organisations can aggregate enough votes to make the cut off percentage. Moreover, presently many of the members of these organisations do not vote for their candidates because they do not want their vote to go waste in support of a candidate who is destined to lose. But since in a proportional representation system their votes will count so the supporters of mass movements will vote for their candidates. Indeed in Germany the Green Party has made a mark because of PR. Even though given the overall capitalist domination of the polity in this country the mass movements will not be able to bring about any substantial change in the way this country is run nevertheless they will at least have more of a voice in legislatures and the parliarment than they do now.

  10. FPTP might be unfair to the loser but proportional representation will result in more hung houses.

    while there are benefits to this scheme or that in the process of electing representatives, i think the bigger problem of democracy is responsiveness of public officials to public welfare during their term in office.

    in the institutional design of democracy the idea of elected representatives was that they would influence those in power away from arbitrariness and abuse of power additionally with the government machinery to be headed by a ‘winning group’ of representatives for a term. but those representatives themselves having become professional dynasties as part of the party-based system, the link with public welfare has weakened and so has the influence away from abuse of power.

  11. i guess the main point of the above analysis is that the winning party should be a bit more humble since the ‘mandate’ is not as clearly for the winners as they and the media would like to project. In fact, twice as many people voted against the SP than for it, in the case of UP.

    To make it an argument for proportional representation, misses the fact that political parties are very much like mafia organisations based on a dominating individual or clique on top operating not so much for ideology or vision but merely realpolitic. In the US, say, we see that even the candidates are openly elected by party members, but in India they are almost always arbitrarily chosen by the ‘high-command’ which itself almost never stands for open elections. So in the case of PR, the public would even lose the knowledge of who exactly their own representative is, since the party high-commands would just appoint whoever they like even more than they already do.

    As a student i used to wonder how, if someone, say, became a doctor with first division marks, he or she is to be regarded as a ‘good’ doctor – clearly if he/she got 60% marks he/she is wrong at least half the time that he/she is right. And even 3rd division doctors are out there treating patients who according to their marks are (like the SP in UP) wrong twice as many times as they are right.

    My point here is again the the key in ‘good’ democracy is design of institutions to make whoever is elected to function without (i add a qualifier) too much arbitrariness or abuse of power. this form of democracy i argue for maybe, i admit, rather limited tho possibly more difficult to achieve involving continuous struggle, including what we see on this website, rather than merely the discrete game of elections.

  12. In a representative democracy if the majority of population is dishonest , uneducated and clannish in nature than the majority of representatives should be too. Seems lil representativee democracy is working perfectluy in India

We look forward to your comments. Comments are subject to moderation as per our comments policy. They may take some time to appear.

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s