A farmer's two percent
Published 1 October 2026
On paper, the Pradhan Mantri Fasal Bima Yojana is the most noble product in this file: crop insurance for the farmer, subsidized by the taxpayer, priced so that a peasant pays two percent and the state pays the rest. Launched in 2016, it is the largest crop insurance scheme on earth. It is also, by the arithmetic tabled in Parliament and reported from the states, one of the most profitable insurance books ever underwritten in India. The profit did not come from the farmer's two percent. It came from the gap between what the exchequer paid in and what the fields were paid out.
The national ledger was laid out in the press from government data: in the scheme's first five-plus years, insurers collected about Rs.1.59 lakh crore in premiums - mostly central and state subsidy - and paid about Rs.1.19 lakh crore in claims, pocketing a margin of roughly Rs.40,000 crore. A quarter of everything that flowed through the scheme stayed with the insurers. Crop insurance is supposed to be the thinnest-margin, highest-social-purpose line in the business; private insurers bid for state clusters precisely because, drought years aside, it has been a money pump.
The state-level numbers show the pump at full pressure. From Union agriculture ministry data for Haryana, insurers collected Rs.2,827 crore in premiums over 2023-25 and paid Rs.731 crore in claims - a profit of Rs.2,096 crore, a claims ratio of about 26 paise on the rupee. In a scheme where the farmer's premium is capped at 1.5-2% and the rest is public money, three of every four rupees the taxpayer routed through Haryana's crop insurance never reached a field. When the state-level numbers surfaced, the political argument that followed - a former state finance minister accusing insurers of looting farmers - named the farmer as the loser. The fuller truth is worse: the farmer was undercompensated and the taxpayer was overcharged, and the same margin was booked twice.
How a welfare scheme becomes a margin
The design failure is not hidden; it is structural, and everyone in the chain can name it. Insurers bid for a state's cluster by quoting a premium rate on the sum insured; the bid is priced against actuarial models of rainfall and yield that the insurer understands and the state agriculture department mostly does not. Bad years trigger claims; normal years - and most years are normal - trigger pure profit, with no mechanism that returns the surplus to the exchequer that funded it. Add the claims frictions every farmer knows - yield data disputes, delayed assessments, units carved so that a village's loss averages away - and the scheme's protection runs thin exactly where its premium runs thick.
The farmer's experience of the gap is not abstract. His premium is deducted at the loan window whether he wants the policy or not, because the scheme rides on the crop loan. His claim, when the rains fail, arrives after the sowing he needed it for. And his compensation is computed on cluster yields that can declare his ruined field part of a normal block. The industry's defenders reply, correctly, that severe years reverse the arithmetic - insurers have taken losses in drought clusters - and that without profit no private capacity would enter the scheme at all. Both are true, and neither answers the question the Haryana numbers ask: a quarter of the national book, retained. Welfare schemes are allowed to cost money; that is what makes them welfare. What India built instead is a welfare scheme that pays like a casino on most years and calls it agriculture policy. The premium machine did not need to capture this one. It was handed to it, bid by bid, by the design.
The yield-cut clause
The scheme's mechanics repay a closer look, because the margin is engineered in the details rather than taken at the door. Claims under PMFBY are assessed not on the individual farmer's field but on the "insurance unit" - a block of villages whose average yield decides everyone in it. A farmer whose crop failed on his own land collects nothing if the block's average held; a farmer in a failing block collects whether or not his field failed. The design exists for a defensible reason - individual farm-level assessment across crores of smallholdings is impossible - but its effect is to convert insurance from protection against my loss into a lottery on my neighbors' yields, and it hands the insurer a claims lever that runs on administrative choices: how the units are drawn, whose crop-cutting experiments count, when the data arrives. Delay is the quietest lever of all. A claim settled after the next sowing has failed is not protection; it is compensation archaeology.
The states, to their credit, have voted on the scheme with their feet at various times - several large ones have exited or restructured their participation, citing precisely the arithmetic this chapter describes, and the Centre has reworked the scheme's terms more than once: capping insurer premiums in some crops, tightening claim timelines, threatening penalties for delay. The scheme survives because the alternative is worse - an uninsured farmer is one failed monsoon from the moneylender, and everyone in the debate knows it. That is what makes the Rs.40,000-crore margin so corrosive: it taxes the one instrument standing between the Indian farmer and ruin, and it does so with public money, behind a welfare brand. The premium machine's most profitable book is also its most protected from criticism, because criticizing it sounds like criticizing the farmer. It is the opposite. It is billing the farmer's protector and pocketing the change.
The scheme that succeeded too well
None of this was the plan. When PMFBY launched in 2016, it replaced a patchwork of failed crop schemes with a design that looked, on paper, like a breakthrough: technology-based yield assessment, remote-sensing pilots, a capped farmer premium of 1.5-2%, and private insurers competing for state clusters to bring discipline the old state-run scheme never had. Enrollment exploded into the crores; the scheme became the largest of its kind anywhere; and for the first time, crop insurance in India was a real market with real capacity. That is exactly what makes the profit ledger so corrosive. The scheme proved that Indian agriculture could be insured at scale - and then let the margin on that proof be set by the side of the table with the actuaries, against the side with the agriculture department.
The states' retreat since then is the market's own verdict. Faced with premium bills in the thousands of crores and claim flows that did not match their drought years, several states restructured, renegotiated, or stepped back from the standard design, and the Centre has amended the scheme repeatedly - voluntary enrollment for loanees, district-level reckoning, penalty clauses for delayed claims. Each amendment is an admission that the original pricing trusted the wrong party. The lesson generalizes beyond agriculture, and it is this file's lesson in miniature: wherever one side of an insurance contract understands the risk and the other side merely funds it, the funding side pays for the understanding. A farmer's two percent and a taxpayer's ninety-eight are the same transaction seen from two windows, and the machine collects at both.