The Most Expensive Word in Indian Politics

When a star performer leaves a company, the company sees it as a loss. Someone quietly asks what she knew, what she saw coming, and what the competitor offered that the company did not. These questions may be uncomfortable, but they are useful.

When a leader leaves a political party in India, the party holds a press conference. By the end of the day, he has a new name: gaddar, Mir Jafar, Vibhishan, opportunist, agent. No one asks what he knew, because the answer has already been given: he is a traitor, and traitors know nothing worth hearing.

I recently wrote about leaders who treat every criticism as an attack. This is a quieter form of the same deafness. Criticism comes in words, but exit shows up in actions: a resignation, an ally walking out, a worker who stops showing up. It is the most honest feedback a party can get, and Indian politics has built an entire vocabulary for refusing to read it.

The word that ends the inquiry

Every party has its word, and every party uses it.

In the run-up to the 2021 Bengal election, as leader after leader crossed from the Trinamool to the BJP, Mamata Banerjee called them Mir Jafars, after the general whose betrayal lost Bengal at Plassey, and said the party was better off without them. When the TDP walked out of the NDA in 2018, BJP president Amit Shah wrote to Chandrababu Naidu that the decision was unilateral and driven wholly by political considerations rather than concern for Andhra Pradesh. When Jyotiraditya Scindia walked out of the Congress in March 2020, the party expelled him for anti-party activities. A year later, Rahul Gandhi said Scindia could have become Chief Minister had he stayed, and would only ever be a backbencher in the BJP. Scindia's reply was a single line: he wished the same concern had been shown while he was still in the Congress.

Three parties, three different situations, the same response. The departure is turned from a message into a motive. Once the motive is named (greed, ambition, betrayal, politics), there is nothing left to examine. The label does not explain the exit. It closes the file on it.

That is why "gaddar" is the most expensive word in Indian politics. It costs nothing to say and feels righteous, but it costs the party the only free piece of market research it will ever get.

Leaving is the first warning

More than fifty years ago, the economist Albert O. Hirschman noticed that when an organisation starts to decline, people respond in one of two ways: they complain, or they leave. Complaints tell you what is wrong. Leaving only tells you that something is wrong, but it tells you first, and it tells you loudly. He also saw that the first to leave are usually the ones who care most about quality, sense problems early, and have better options elsewhere. The organisation loses its best diagnosticians at the worst possible time.

Viewed this way, the pattern in political parties is troubling. The leader who leaves a year and a half before an election is rarely the deadwood. Deadwood stays; it has nowhere else to go. Early leavers are usually the ones who can win elsewhere, so their judgement about the party's future matters most. When Scindia left, twenty-two MLAs followed, and the Kamal Nath government fell ten days later.

The silence in the room

Some exits never make the news, and I have watched that one most often.

I have attended more review meetings than I can remember. The subject changes: a minister's performance, an MP's standing, an MLA's chances. The room does not. Everyone in it knows the truth. The district office-bearers know, the local workers know, and even those who travelled in from the constituency that morning know. The man is unpopular in his own seat. But no one says it. Not in front of the leader, not when the man's patron is present, and definitely not after watching the last person who spoke up being quietly pushed aside.

So everyone in the room agrees that the situation is under control. The report goes up the chain. The workers go home, and some of them quietly stop canvassing the streets they have worked for twenty years. They don't resign. They don't join another party. They just stop.

That is exit without leaving. Those who built on Hirschman’s work saw something else: loyalty can delay people from leaving, but it can also make them silent. In Indian politics, that silence runs on an industrial scale, because speaking up is read as disloyalty and leaving as betrayal. The worker who won't betray and isn't allowed to complain does the only thing left. He goes quiet. And the party, which counts members by the register and not by their effort, notices only when the votes are counted.

All exits are not prophecies

Not every departure is a verdict on the party. Sometimes the people who leave are wrong.

Many leaders who crossed to the BJP before Bengal 2021 read the wind and misread it. The Trinamool won comfortably, and within weeks the tide turned. Mukul Roy, who had joined the BJP years earlier and risen to its national vice-presidency, was back in the Trinamool, and others were knocking. The TDP's exit from the NDA in 2018 was a bet against the BJP; a year later, the TDP was routed at home while the BJP won a bigger national majority, and in 2024, the TDP returned to the NDA.

Sometimes the people who leave are right. The Trinamool walked out of the UPA in September 2012. The DMK followed in March 2013. In 2014, the Congress fell to 44 seats. After leaving the Congress in 2020, Scindia became a Union minister, and the BJP won Madhya Pradesh comfortably in 2023.

Maharashtra offers the strangest proof of all. In 2022, Eknath Shinde did not leave the Shiv Sena; he took most of its MLAs with him, and the Election Commission eventually gave his faction the party's name and symbol. In 2023, Ajit Pawar did the same to the NCP. The side that was left behind built its campaign around a single word: gaddar. In the June 2024 Lok Sabha election, the departures looked like a mistake. Uddhav Thackeray's Sena won nine seats to Shinde's seven, and Sharad Pawar's NCP won eight to Ajit Pawar's one. Five months later, in the assembly election, the same state reversed the verdict: fifty-seven seats to twenty, forty-one to ten. The same exits, read two opposite ways within five months.

Notice, too, what happens to the label. When an exit is big enough, the people who leave keep the party's name, its symbol, and the office, and "traitor" becomes a word only the smaller side still uses.

So the real skill is neither to obey every exit nor to dismiss them all. It is to read them. Who left? How capable were they? Where did they go? Did they leave with a press conference and a new party scarf, or did they stop asking for a ticket? Most importantly, did the people who stayed go quiet?

In the past decade, most major exits in Indian politics have flowed towards the BJP. That is exactly why the parties on the losing end of that flow can least afford to stop reading them. Bengal also shows how quickly the flow can reverse when the bet goes wrong.

There is a trap here that needs to be named. Sometimes calling defectors traitors turns out to be right. When that happens, the party learns that the label works, and the next time someone leaves, it does not even pause to ask. A party proven right once about its traitors is in more danger than one that never was.

Allies are the market

If individual departures are signals, allies are a market.

An ally knows more than any voter and has more at stake than any commentator. It sees how you share seats, how you treat junior partners, how you behave in private meetings. It also knows what its own voters think of you. When an ally starts to drift, it is pricing your future. Like any market, it can get the price wrong; the TDP in 2018 is the proof. But even a wrong market tells you what the best-informed people believe, and that is worth knowing.

A party that treats its allies like a market learns something either way. A party that treats every drift as a betrayal learns nothing either way.

The competitor you were not watching

There is one more thing parties miss, and it is the most expensive of all.

A war room's competitive intelligence focuses on other parties: their speeches, ads, candidate lists, alliances, and leaders' movements. It does this well. But it is almost completely blind to the challenger who does not yet have a party office.

On 29 March 1982, a film star in Andhra Pradesh started a new party. Nine months later, it won 201 of 294 seats and ended more than a quarter of a century of Congress rule in the state. The ruling party had spent those nine months watching only the opponents it already knew. The pattern keeps repeating in new forms: a movement that becomes a party, a film star whose new outfit takes a slice of the vote at its first outing, and, as I have written before, a cartoon cockroach with no office at all that still managed to force out a Union minister.

The real challenge rarely looks like a party until it becomes one. By then, it is too late to treat it as a rumour. Every war room should ask not only "what are our rivals doing?" but also "who is not a party yet, and is attaching itself to the grievance we are ignoring?"

The morning someone leaves

So what should a party that understands all this actually do?

It holds the press conference later. Before anyone reaches for a word, someone in a quiet room asks the only question that matters: what did they know that we don't?

It looks at the quality and direction of exits, not their number. Ten loud departures by people going nowhere matter less than one quiet exit by someone who could have won their seat, and less again than the aspirant who stops asking for a ticket.

It listens for the exits that never make the news: the worker who still shows up but no longer canvasses, the family that quietly votes for someone else, the meeting where everyone agrees too quickly.

It treats its allies as a market, not a loyalty test. It asks what they are betting on. It neither obeys them nor insults them.

It widens its list of competitors to include what is not yet a party, and gives someone the job of watching it.

And it makes sure bad news has at least one road to the top that does not run through the people the news is about.

In my view, a party that does only the first of these, that waits a day before calling anyone a traitor, will learn more about its own position than from any survey it commissions this year.

Every exit sends a message. The party is always the last to read it. Not because the message is hidden, but because the party is the only reader that has already decided what it says.