8 minutes reading time
Every time someone says "India should try ranked-choice voting," I picture a beautifully engineered irrigation canal being dug through the Sahara. The engineering is impeccable. The location is the problem.
Ranked-choice voting (specifically the variant called Single Transferable Vote, or STV) is a genuinely clever system. I want to be fair to it before I explain why it would struggle badly in India. Because the cleverness is real. The problem is us, not the concept.
In STV, instead of one constituency electing one MP by plurality, you merge several constituencies into a larger multi-seat district. Say five constituencies become one district electing five MPs.
Instead of just marking an X next to a name, you rank candidates: 1, 2, 3, 4, 5. Your first preference counts first. If your candidate gets more than enough votes to win a seat, their surplus votes get transferred to your second preference. If your candidate is eliminated because they're last, your vote transfers entirely to your second preference.
The process continues in rounds until all five seats are filled. Almost every ballot ends up contributing to electing someone, unlike FPTP where 40% of votes can vanish completely.
The result: a five-seat district where parties win seats roughly proportional to how many people supported them. Ireland has used this since 1921. Malta uses it. Bits of Australia use it for the Senate. It works. In the right conditions.
The right conditions being the key phrase.
India votes by symbol.
This is not a minor detail. This is fundamental to how a billion-person democracy with a literacy gap functions. The lotus is BJP. The hand is Congress. The cycle is Samajwadi Party. Voters who can't read a single word of any language can walk into a booth and correctly vote for their candidate because they recognise the image.
This system took decades to build. It's how you run a functioning election across Tamil Nadu, Bihar, Assam, and Nagaland simultaneously, where the electorate speaks dozens of languages and has wildly varying literacy levels.
Now. Under STV, you'd merge five current constituencies into one district. That district might have 25 to 40 candidates running. Each voter needs to rank at least several of them. Even if you print symbols on the ballot, you now have forty symbols crammed onto one sheet. Voters who rely entirely on visual recognition now have to find their symbol among forty, assign it a rank number, and do the same for their second preference, their third.
For a large chunk of India's electorate (particularly in rural areas, particularly first-generation voters, particularly older voters), this is not a simplification of democracy. It's a barrier to it.
Imagine a constituency that elects five members to Parliament, not just one. Five seats available in a single district.
Instead of putting a cross next to one name, you rank candidates: 1, 2, 3, 4, 5.
A candidate needs a specific "quota" of votes to win a seat (calculated as total votes divided by six, plus one). If a candidate gets more than the quota on first-preference votes, they're elected. Their excess votes are then transferred to the second-choice candidates marked on those ballots.
If nobody reaches the quota, the candidate with the fewest first-preference votes is eliminated, and their ballots transfer to those voters' second choices.
The process continues in rounds until all five seats are filled. Almost every ballot ends up contributing to electing someone, unlike FPTP where 40% of votes can vanish completely.
It sounds great on paper. In practice, India poses three severe challenges:
Under STV, you don't just pick a symbol. You rank candidates in order of preference. That requires reading candidate names, understanding the ordinal ranking logic, and filling out numbers (1, 2, 3...) accurately.
For a large chunk of India's electorate (particularly in rural areas, particularly first-generation voters, particularly older voters), this is not a simplification of democracy. It's a barrier to it.
India built its entire voting infrastructure (from paper ballots with party symbols in 1951 to modern EVMs) around a single, low-friction action: press a button next to a picture. Switch to STV and you replace that single button press with a multi-step numerical ranking task.
The risk of ballot invalidation (voters putting "1" next to three names, or skipping numbers, or leaving half the ballot blank) is enormous.
When a voter stops ranking (marking preferences 1 and 2 while leaving the rest blank) and their top two choices are both eliminated, their ballot becomes "exhausted." It no longer counts toward any candidate. In countries with simple ballots and educated electorates, truncation is a manageable minority problem. You lose maybe 5-10% of votes to exhaustion.
In India, with a complex multi-candidate ballot and voters who are accustomed to marking exactly one symbol, the exhaustion rate would likely be much higher. Voters who rank only their first choice (because that's all they understand how to do) produce exhausted ballots at scale. And an exhausted ballot under STV is the same as a wasted vote under FPTP.
STV requires multi-member constituencies (districts that elect 3, 4, 5, or more representatives). That means merging existing single-member constituencies into giant super-constituencies.
Instead of one MP for Coimbatore North, you'd have five MPs for the whole of Coimbatore district.
That destroys the local MP connection. In a country where citizens rely on their local MP for everything from water supply complaints to police station disputes, replacing a local representative with a panel of five district-wide MPs makes accountability fuzzy. Who do you complain to when the road breaks?
And redrawing boundaries in India is politically explosive. The delimitation freeze (which I'll write about separately) shows exactly how much tension this creates. Asking the country to redraw 543 constituencies into 108 is not a reform. It's an earthquake.
Supporters of STV say it eliminates strategic voting. It doesn't. It just changes the strategy.
Instead of "don't vote for the third party," the strategy becomes "how do we instruct our voters to transfer their preferences?" Parties produce "how to vote" cards, telling supporters to rank Candidate A first, Candidate B second, Candidate C third.
Done messily (which is how politics actually works), it splits the vote internally and costs seats. Parties will agonise over how many candidates to field. They'll make deals: "Rank our candidate second and we'll rank yours second in the next district." Transfer alliances. Pre-poll arrangements about preference orders.
Kerala already has an informal version of this between the Left Democratic Front and the United Democratic Front, alliances built partly on where each bloc's supporters should direct their second preferences. Under STV, this becomes formalised, game-theoretic, and incomprehensible to anyone not following it closely.
Here's the funny thing: India already uses STV. Just not for the Lok Sabha.
The Rajya Sabha (the Upper House) is elected by Members of the Legislative Assembly using STV. The voters are MLAs, who are literate (compared to the general public), politically sophisticated, and can be briefed in detail on the counting process.
Legislative Councils in states like Tamil Nadu and Maharashtra (when they existed) used STV for graduates' and teachers' constituencies.
STV works when the voters are equipped to use it. The question for India isn't whether STV is theoretically better than FPTP: in many ways it is. The question is whether you can operate it at the scale of a billion general voters across twenty-two official languages and wildly varying literacy rates.
The answer right now is: probably not without massive disruption, a long transition period, and a voter education campaign that would itself need to be one of the largest public information operations in Indian history.
STV is not a fraud. In the right setting, it genuinely improves on FPTP. It's proportional, it reduces wasted votes, it lets voters choose between candidates within a party rather than just accepting whoever the party boss put on the ticket.
But India's "right setting" is not a five-seat STV district with forty candidates and a ballot you need to rank. Not yet. Not without first solving the literacy gap, the symbol dependency, the multilingual ballot problem, and the constituency map.
I’ve been thinking a lot about MMP since I wrote those notes the other day. It doesn't require any of that. It keeps the current single-member constituency structure intact. It adds one extra vote on the same ballot. It uses the infrastructure that already exists.
STV is the better ideal, maybe. MMP is the achievable reform.
And a reform you can't implement is just a fantasy you're fond of.
A beautiful irrigation canal in the desert is still a desert.