After a Deficient Monsoon, the Insurance Question Is District by District: Reading Kharif 2026 Before Claims Settle

The southwest monsoon has closed its account for 2026, and for anyone carrying Kharif risk — a crop-loan book, an insurance portfolio, a procurement plan — the headline is the least useful part of it. India as a whole was short. The districts that will generate claims, repayment stress and supply shortfalls were short by very different amounts, at very different points in the season, and with very different insurance cover standing behind them. October is the month to read that district picture properly, before yield estimates harden into settlements.

What the season actually delivered

The India Meteorological Department closed the June–September season at 759.4 mm of rainfall against a long-period average of 868.6 mm — 87% of the average, the fourth-lowest total since 2001 and the weakest season since 2015. The monthly pattern matters as much as the total for crop outcomes: June ran about 35% below normal, July recovered to roughly normal, August slipped back to about 16% below, and September closed about 8% below. A crop that was sown late after a dry June and then stressed again in August has not experienced the same season as the 87% national figure suggests.

The regional split is sharper still. East and Northeast India received 1,017.1 mm against a normal of 1,367.3 mm, and South Peninsular India 542 mm against 716.2 mm, while Central India was only about 3% below normal and the Northwest about 6% below. Of 36 meteorological subdivisions, 18 were normal, 17 were deficient and one — Odisha — recorded excess rainfall. At district level, 313 of 741 districts ended the season deficient or large deficient, meaning at least 20% below normal; 217 of them received less than 70% of normal rainfall, and 71 received only about half.

For a portfolio, that distribution is the finding. A lender or insurer whose exposure sits mainly in Central India is looking at a broadly normal season with pockets of stress. Exposure concentrated in the South Peninsula or the East and Northeast is looking at a materially different loss year — from the same national headline.

Timing decides which insurance trigger is even relevant

A seasonal deficit does not convert into a single kind of loss. A dry June raises prevented or delayed sowing and forces re-sowing costs before a crop is established. A dry August, at flowering and grain-filling for many Kharif crops, cuts yield on a crop that was successfully established. Localised events — a hailstorm, a cloudburst, inundation in one block — can destroy a crop in a district whose seasonal total looks acceptable.

That distinction is built into the public insurance design, and risk teams should use the same separation in their own books. Under the Pradhan Mantri Fasal Bima Yojana, prevented sowing, area-based yield loss on standing crops, localised calamities assessed at farm level, and specified post-harvest losses are different covers with different evidence and different settlement paths. Treating "monsoon was deficient, therefore claims will be high" as a single assumption mixes those paths together and will misstate both the size and the timing of recovery on an insured book.

The cover that was in place before the rain failed

Insurance only changes a lender's or buyer's outcome if it was bought before the loss, in the district where the loss occurs. On that measure, Kharif 2026 entered the season with more cover than the year before. The government's August backgrounder records PMFBY operating in 25 states and Union Territories in Kharif 2026, with 241.38 lakh farmers insured over 278.12 lakh hectares as of 27 August — already above Kharif 2025's 229.77 lakh farmers and 269.38 lakh hectares. The same release puts the scheme's longer record at more than 92.46 crore farmer applications insured since Kharif 2016, with claims above ₹2.06 lakh crore paid, and a ₹12,200 crore allocation for 2026–27. West Bengal's return to the scheme from Kharif 2026 is part of this season's coverage expansion; Bihar is recorded as returning from Rabi 2026–27.

Read alongside the weather-based scheme, the insured footprint is larger. Government data reported for PMFBY together with the Restructured Weather-Based Crop Insurance Scheme put Kharif 2026 at 10.6 crore applications covering 29.7 million hectares, with sum insured of ₹1.86 lakh crore, up from 7.17 crore applications and ₹1.45 lakh crore in Kharif 2025. The two sets of figures are not interchangeable — one is the PMFBY farmer and hectare count in the backgrounder, the other includes the weather-based scheme — and a risk team should keep them labelled that way. What they agree on is direction: more area and more sum insured stood behind this Kharif crop than behind last year's, in the season that turned out to need it.

The premium structure explains why participation can move at that scale. Farmers pay a capped share — 2% of sum insured for Kharif crops, 1.5% for Rabi and 5% for commercial and horticultural crops — with the balance shared between Centre and states. For a lender, the operational point is not the subsidy itself. It is that an insured borrower and an uninsured borrower in the same deficient district are different credits from October onwards: one has a defined, if delayed, compensation path, and the other carries the full yield shortfall into repayment.

How the yield number will be produced — and where to challenge it

In a normal year, the method behind the yield estimate attracts little attention. In a deficient year, it decides settlement size across entire insurance units, so its mechanics belong in a risk review now, not after claims are announced.

Two technology systems now sit inside that method. YES-TECH estimates yield using remote sensing and related technology, and its weight in the yield calculation has risen from at least 30% to up to 50% in some states. WINDS builds the underlying weather record from automatic weather stations and rain gauges at block and gram-panchayat level, which is also the data layer weather-based and parametric covers depend on. Crop-cutting experiments, captured digitally, remain part of the actual-yield calculation, and claims are calculated and paid through the scheme's digital portal.

Each element carries a specific limit an institutional user should price in. A technology yield estimate smooths local variation inside an insurance unit; a farm on the dry edge of a unit can lose more than the unit average that determines its payout. Hyperlocal weather data is only as representative as station density and upkeep in that block. And an area-based yield cover, by design, does not pay for a purely localised loss unless that loss is reported and assessed under the localised-calamity route, within its conditions and timelines. None of this makes the cover unreliable. It means the expected recovery rate on an insured portfolio is a district-level estimate with a range, not a national percentage to apply uniformly.

What to do in October, before settlements arrive

Split the book by rainfall outcome, not by state total. Map Kharif exposure to the district rainfall outcome — deficient, large deficient, or normal — and keep the 71 districts at roughly half of normal rainfall as a separate watchlist. That is where repayment stress and claims volume will concentrate, regardless of the state average.

Separate insured from uninsured exposure inside each deficient district. For insured exposure, model the compensation path and its likely timing, including which cover type the dominant loss fits. For uninsured exposure, the same yield shortfall is a direct credit and supply problem, and it needs a different response and an earlier one.

Do not close the weather file at 30 September. The IMD's October outlook is a forecast, not an outcome: it expects October rainfall to be below normal, at less than 85% of the long-period average over most of the country, with October–December rainfall over South Peninsular India most likely normal. Labelled as a forecast, it still matters in October for two live decisions — post-harvest losses on crops still in the field, and the soil-moisture starting point for Rabi lending already covered in this series.

Prepare the evidence before you need it. Districts where your own crop-condition, arrivals or ground reports diverge from the emerging yield picture are the districts where representation through the claim process is most valuable. That comparison is only possible if the institution's own signals were recorded during the season, not reconstructed after settlement.

Where Farmioc fits

Farmioc DataStore organises supply and demand, yield and harvest outlooks, and satellite and ground-based data by crop and region for the financial sector, agribusiness and trading teams. Joined to a lender's or insurer's own book, that is the layer that turns a national monsoon total into the view this season actually requires: which districts were short, when in the crop cycle they were short, what cover stands behind the exposure there, and where the institution's own signals agree or disagree with the yield estimate that will drive settlement.

A deficient monsoon is not, by itself, a portfolio loss. It becomes one — or does not — district by district, cover by cover, and estimate by estimate. October is when that reading is still cheap to do. By settlement, it is only an explanation.

Farmioc is India's agriculture data analytics and agri trading company. Explore DataStore for supply, demand and yield analytics, or PriceHub for commodity market intelligence across 200+ crops.