NCDEX Is Selling Chennai Rainfall Risk: Weather Derivatives Against Parametric Insurance for an Indian CFO

(Re)in Asia reported on 4 September 2026 that NCDEX has extended its weather derivatives to hedge Chennai rainfall risk. It is the exchange's second weather contract after Mumbai, and it covers both of the country's major monsoon seasons, the southwest monsoon and the northeast monsoon that delivers the bulk of Chennai's annual rainfall between October and December.

That second season is why the contract matters commercially. Chennai's severe rainfall events cluster in the northeast monsoon, and the loss record is not theoretical. General insurers received claims of over Rs 2,500 crore from the Tamil Nadu rains of 2015 (Business Standard, 9 December 2015), and fire and engineering covers accounted for the highest claim values in those floods (Business Standard, 11 December 2015). The industrial belt around Sriperumbudur, Oragadam and Maraimalai Nagar sits inside the same rainfall regime.

The cost trend has not improved. Insurers expect claims of nearly Rs 4,000 to 5,000 crore from the 2026 Gujarat rains, mainly on property lines, against Rs 1,500 to 2,000 crore from the 2024 Gujarat floods (Business Standard, 2 August 2026). Rainfall loss cost is repricing property programmes, flood deductibles are moving up, and treasury teams are being asked what they can hedge outside the insurance market.

Until now the answer for an Indian corporate was a parametric policy placed with a domestic insurer or through a GIFT City vehicle. The Chennai contract gives a second answer. Both pay on a measured weather index rather than on a surveyed loss, so they look similar from a distance. The differences show up in law, accounting, tax and treasury operations, and they decide which instrument belongs in which part of the programme.

A parametric policy is an insurance contract. It is issued by an IRDAI-registered insurer, it sits under Indian insurance law, and it requires insurable interest. The payout formula references an index, and the buyer still has to be exposed to loss from the insured event. That link back to an underlying economic loss is what keeps the product inside the definition of insurance rather than a wager.

An exchange-traded weather derivative is a different object: a contract on a SEBI-regulated exchange, settled in cash against a published index at expiry, with no insurable-interest test, no proof of loss, no surveyor and no claim. If the index breaches the strike, the contract pays whether the buyer suffered a rupee of damage or not. If the buyer suffered heavy damage and the index did not breach, it pays nothing and there is no argument to be had.

The two instruments therefore cover different questions. The parametric policy answers "I will suffer a loss from heavy rainfall and I want it indemnified quickly without a claims process." The derivative answers "my revenue, procurement cost or output volume is statistically sensitive to rainfall and I want that sensitivity flattened." A treasurer who wants the second thing has, for the first time in India, an instrument built for it.

Both instruments carry the same four families of basis risk. Spatial basis risk is the gap between the settlement station and the site. Temporal basis risk is the gap between the measurement window and the period in which the loss accrues. Metric basis risk is the gap between what the index measures, typically cumulative or daily rainfall depth, and what causes the loss, typically inundation depth, drainage failure or river discharge. Demand basis risk is the gap between physical rainfall and the revenue effect it produces. What differs is your ability to do anything about it.

On a parametric policy you negotiate the wording. You can move the reference grid cell, add a second station, change from a daily maximum to a three-day cumulative, insert a step payout ladder in place of a binary trigger, and align the risk period with your fiscal calendar. Each adjustment is a repricing conversation with the underwriter, and each one reduces basis risk against your specific site.

On an exchange contract the specification is fixed. The settlement station, the accumulation window, the tick size and the expiry are what make the contract fungible, and fungibility creates liquidity. You take the specification as written or you do not trade, and a plant 40 km inland from the settlement station keeps whatever spatial mismatch that distance produces.

Before either instrument goes to the board, back-test the proposed index against your own loss and revenue record over at least 15 to 20 years of India Meteorological Department data for the settlement station. Count the years the index would have paid when you had no loss, and the years you had a loss and the index would have paid nothing. Those two counts are the honest description of the hedge and belong in the placement memo. The same discipline applies to any index cover, as set out in our note on parametric trigger design and basis risk.

The two instruments diverge most sharply in the finance function, and it is usually discovered late.

The derivative

A weather derivative is measured at fair value through profit or loss by default, so every reporting date the change in fair value hits the income statement. If the exposure it hedges is a forecast procurement cost or a revenue shortfall two quarters out, the hedge introduces earnings volatility in the interim, the opposite of why it was bought.

Hedge accounting under Ind AS 109 removes that mismatch and is not automatic. It requires formal designation and documentation at inception naming the hedging instrument, the hedged item, the nature of the risk hedged, and the method of assessing effectiveness. It requires an economic relationship, credit risk that does not dominate value changes, and a hedge ratio consistent with actual quantities. Where the hedged item is non-financial, such as forecast sales volume or procurement cost, the rainfall risk component has to be separately identifiable and reliably measurable. That condition is the one weather hedges most often fail, because the rainfall component of a revenue line is a statistical association rather than a contractually specified price component.

The parametric policy

An insurance premium is expensed over the policy period with no mark-to-market. The recovery is a different problem: a claim is treated as a contingent asset and recognised only when realisation is virtually certain, which in practice means when the insurer has admitted it. A parametric policy improves that timing, since the trigger is objective and admission follows publication of the index, though the accrual still lags the loss event.

Get both treatments confirmed in writing by the audit partner before the first trade or the first placement, not at year-end close.

The cost stacks are built differently, and a straight premium-versus-price comparison understates the gap.

On the insurance side, a commercial general insurance premium attracts GST at 18 percent on the full premium, with input tax credit generally available where the cover relates to business use, subject to the blocked-credit list in Section 17(5) of the CGST Act. The GST exemption granted in September 2025 applied to individual life and health policies; commercial property, engineering and parametric covers bought by a company remain taxable.

On the derivative side, the transaction value sits outside the GST base. GST at 18 percent applies to brokerage and exchange transaction charges, a fraction of the notional, and stamp duty is levied on the transaction at the uniform rates collected by the clearing corporation. On indirect cost alone, the exchange route is cheaper per rupee of notional.

Direct tax is where the caution belongs. An insurance premium is deductible business expenditure and a recovery is taxable as business income, both settled ground. Section 43(5) of the Income-tax Act defines a speculative transaction as one settled otherwise than by actual delivery, with carve-outs including hedging transactions and eligible transactions in commodity derivatives on a recognised association. Whether a cash-settled weather contract falls inside a carve-out is a question of characterisation that has not been tested for this product.

The parametric policy is a single cash outflow. You pay the premium at inception and then hold a credit exposure to the insurer for the life of the cover. That exposure is regulated: an Indian insurer must maintain a solvency ratio of at least 150 percent of the required solvency margin, and the retention is typically reinsured. There is no collateral to post and no daily process to run.

The exchange contract works the other way. The clearing corporation novates the trade, so the buyer holds no exposure to an individual counterparty. The cost is cash mechanics:

1. Initial margin is posted before the position opens and is sized by the exchange, not by the buyer.
2. Variation margin moves daily with the mark, so a position out of the money mid-season calls cash while the rainfall season is still running.
3. Position limits and available open interest cap the notional you can put on. A newly listed single-city contract has thin liquidity, and a treasury needing size may not fill it.
4. Exiting before expiry requires someone to take the other side at a price you accept.

The point a treasurer has to plan for is that the hedge which pays at expiry can still call cash from the treasury in the weeks before the event it protects against. Size the margin line at the same time you size the hedge, and agree in advance who is authorised to fund a call mid-season.

Settlement carries no loss adjuster, no surveyor appointment, no document schedule and no negotiation. The contract settles against the published index at expiry and the money moves through the clearing member. A parametric policy comes close, with the insurer still verifying the data source and admitting the trigger, which in Indian placements has typically run days to a few weeks after the index is published.

The question that settles most of these debates is not which instrument is cheaper. It is which one counts.

Lenders count policies. Facility agreements and hypothecation documents name required insurances, set minimum insurer ratings, require the lender as loss payee or mortgagee, and require assignment of policy proceeds. A cash-settled exchange contract cannot be assigned in that form, and the payout is a treasury receipt rather than a reinstatement of the secured asset. A rainfall hedge does not discharge an insurance covenant.

Auditors count documentation. A derivative without Ind AS 109 designation at inception is a fair-value position, however clearly it was bought as a hedge. Designation is a day-one document or it does not exist.

Boards count authorisation. Listed entities in the top 1,000 by market capitalisation must maintain a risk management committee under SEBI's LODR Regulations, and a treasury starting to trade a new instrument class needs a hedging policy approved by that committee before it opens a position. The policy should name permitted instruments, maximum notional, maximum tenor, the margin line, who may execute, and how the position is reported each quarter. An annual insurance renewal creates none of this machinery, a derivative programme creates all of it, and the running cost is part of the price.

The reverse point is equally practical. A property insurance programme with flood cover, machinery breakdown and business interruption responds to the damage and downtime that follow a Chennai-scale rainfall event, which is what the 2015 claims record shows. The rainfall index contract responds to rainfall. Only one of them rebuilds a switchgear room.

The allocation follows the exposure rather than the price.

1. Physical damage and downtime at a named site stay on the insurance programme. Fire, engineering and business interruption cover with a rated flood section is the primary instrument, and the 2015 Tamil Nadu claims profile is the evidence for it.
2. Deductible and time-excess exposure suits parametric cover. Where the property programme carries a heavy flood deductible or a 7 to 14 day time excess on business interruption, a parametric layer sized to that gap pays quickly and without adjustment, the use case set out in our note on parametric rainfall cover for corporates.
3. Revenue, footfall and volume sensitivity suit the exchange contract. Retail chains, contractors losing working days, cement and paint companies with monsoon demand troughs, and hospitality operators carry rainfall-linked revenue variance that will not clear an insurable-interest test.
4. Size decides the venue. Liquidity in a newly listed city contract caps notional well below what a large treasury needs. Where the required cover runs into hundreds of crores, a parametric placement with domestic capacity or a GIFT City vehicle carries size the order book cannot.
5. Run the two together where both fit. A frequency layer on the exchange, rolled each season, below a parametric or indemnity layer for severity is a defensible structure for a Chennai or Mumbai exposure. Document both as one programme so the board sees one retained-risk number.

The test at the end of the analysis is what the instrument does on the worst day. An instrument that pays cash into treasury while the plant is under water and the lender is asking about reinstatement is a treasury hedge and belongs in the treasury policy. An instrument that rebuilds the plant and pays fixed costs while the line is down is insurance and belongs in the insurance programme. Buying one and calling it the other is where Indian corporates will get this wrong over the next two seasons.