As climate risks become harder to predict and insure, India must price catastrophe without leaving people unprotected

The monsoon is rarely just weather in India. It is water for a crop, electricity for a city, traffic on a highway and income for millions of people. This year, it has also offered a glimpse of a more difficult future.

In April, the India Meteorological Department forecast the 2026 southwest monsoon will be deficient and warned that the El Niño weather phenomenon was likely to develop during the season. By September, El Niño, a weather pattern in the Pacific Ocean that impacts the southwest monsoon, had strengthened into a very strong event, with the U.S. National Oceanic and Atmospheric Administration saying it will persist through the northern hemisphere autumn and winter. The monsoon itself has been uneven. June brought a sharp rainfall deficit, July recovered nationally but with large regional differences, and deficits persisted in several parts of the country.

A deficient monsoon does not mean India escapes extreme rainfall events. The same season can leave one place dry and overwhelm another. For a farmer, the consequences may be a failed crop or delayed sowing. For a city, they can mean flooded roads and damaged infrastructure. For a company, they can mean a factory standing idle or a warehouse full of ruined goods. In each case, weather eventually becomes a financial problem. Someone has to pay.

Price of changing climate

India is putting more wealth and infrastructure in places exposed to floods, cyclones, drought and other hazards. Swiss Re, the reinsurance group, estimates that property worth $26 trillion to $29 trillion is exposed to natural catastrophe risks in India. Much of this exposure is concentrated in places where a disaster can disrupt economic activity far beyond the immediate damage. Yet, much of the risk remains uninsured. Swiss Re estimated that 93% of India’s natural catastrophe exposure was uninsured in 2022.

Insurance is supposed to make a large, uncertain loss manageable. A household pays a premium so that an insurer will absorb much of the cost if disaster strikes. A farmer does the same where crop insurance is available. A company insures its factory, machinery and stock. But the insurer has to answer a difficult question before it can sell that protection. How large could the loss be, and how often might it happen?

This is where catastrophe models enter the story. They do not predict the next flood or cyclone. They estimate the likelihood of losses of different sizes. Karen Clark, a pioneer of catastrophe modelling who founded Karen Clark & Company after helping establish the field in the 1980s, says uncertainty increases when models look at increasingly rare events. “One important application of the models is pricing, so if the risk goes up, you would expect premiums to go up to reflect that,” she says.

That insight captures a part of the climate story that is often missed. Climate risk is not only about what happens to the physical world. It is also about what happens to the price of living and doing business in it.

Disaster map getting sharper

The models themselves are changing. Alok Kumar, managing director for insurance at Moody’s, the financial intelligence and risk assessment firm, has spent much of his career in this field, including an earlier association with Risk Management Solutions, one of the companies that helped pioneer catastrophe modelling. Kumar says the most significant change in India’s flood and cyclone modelling has been the move from broad areas of risk towards much more precise assessments of particular locations.

“Models can now better capture intense rainfall, urban drainage, local topography, river systems and building characteristics and how each respond during extreme weather events,” he says. “For insurers, this means moving beyond simply assessing whether a region is exposed to catastrophe to quantifying how those risks vary within the same region.”

The difference is considerable. A city is no longer simply a flood risk area. One neighbourhood may be more exposed than another. A factory beside a river may face a very different risk from one on higher ground. A building with better protection may fare differently from one without it.

RMSI, an Indian risk management company, offers a glimpse of how detailed this has become. Its nationwide flood model uses 80,000 simulated flood events across 28 major river basins and 3,400 sub-basins, with assessments extending across more than 19,000 pin codes. The old risk map painted with broad strokes. The new one is beginning to look more like a street map.

That does not mean insurance prices will automatically rise everywhere. Kumar’s point is more subtle. “This ability to understand risk better should lead to better pricing, not necessarily higher pricing everywhere,” he says.

That is the opportunity. But it is also where the trouble begins. What happens when the new number is worse? If a new model shows that a factory, apartment building or warehouse is more exposed to flooding than previously believed, the risk itself has not suddenly changed. The knowledge of it has.

Aon, an insurance broker and risk advisory firm, found in its 2025 global catastrophe risk management survey that catastrophe models are used most extensively for reinsurance decisions and managing the concentration of risk, while their use in day-to-day underwriting remains less widespread. The distinction matters in India, where models are becoming more sophisticated but their ability to influence the price and availability of insurance depends on how deeply insurers incorporate them into underwriting.

The premium for insurance cover may rise. The customer may have to bear more of the loss. An insurer may decide to take less risk in a particular location. And if insurers themselves find it harder or more expensive to obtain protection against very large losses, capacity can tighten.

Clark offers an important qualification. If insurers can obtain enough reinsurance, or the protection they themselves buy against very large losses, higher risk does not automatically mean insurers will withdraw. But if the increase comes as a surprise, coverage can be reduced in the short term and prices can rise.

That is why Praveen Gupta’s warning matters. Gupta, an industry veteran and former chairman and managing director of Raheja QBE General Insurance, is considerably more sceptical about the industry’s ability to respond. His concern is not confined to the models. He argues that competition for market share has encouraged aggressive pricing, even as the underlying risks have changed. People familiar with the Indian insurance market say that the technology is advancing faster than underwriting discipline and risk awareness in parts of the sector.

“I am seriously concerned about the ability of insurers to cope with and protect us from what is unravelling,” Gupta says. His larger point is that a flood, cyclone or drought does not stay neatly inside the category in which an insurance policy is written. “Catastrophe modelling is not just about physical risk. It is health, life, pensions, agriculture, biodiversity losses, supply chains, and intergenerational risks.”

There is a warning here. Better models may reveal risks that the insurance industry had previously underestimated, but models cannot by themselves make those risks cheaper or easier to bear. As Gupta puts it, “Catastrophe modelling is just another cog in the wheel.”

Paradox of knowing more

This creates a strange problem. Better information can make a risk easier to insure because an insurer can understand it. But the same information can show that the risk is more expensive than anyone had assumed. That is where the protection gap becomes more than an insurance statistic.

Globally, the gap between economic losses from natural disasters and the losses covered by insurance reached $424 billion in 2025, according to Swiss Re. India is at the more exposed end of this problem. Its large uninsured share leaves households, farmers, businesses and governments carrying much of the cost when disasters strike.

The farmer is an important part of this story. Crop insurance is one of the few areas in which insurance against weather-related losses has reached substantial scale. The Sixteenth Finance Commission says the Union and state governments together contribute 85-90% of premiums under the Pradhan Mantri Fasal Bima Yojana. However, farmer penetration was just about 40% in the eight years to 2024-25.

Technology is also changing how some of these risks can be covered. The Finance Commission identifies Nagaland as a pioneer in parametric disaster insurance, with predefined rainfall thresholds triggering payments to protect the state and its infrastructure. The scheme was developed with the state disaster management authority, SBI General Insurance and reinsurers including Munich Re and GIC Re.

Such schemes illustrate both the promise and the limits of insurance in a changing climate. When a weather threshold can trigger a payment, money can reach those affected without waiting for every individual loss to be assessed. But insurance still does not prevent the drought, flood or heat from occurring.

That is where adaptation enters the story. Research by the World Resources Institute, a global policy think tank, examined 320 adaptation and resilience investments across 12 countries and found that each dollar invested could generate more than $10.50 in benefits over 10 years.

The logic is simple. Insurance pays for some of the consequences of risk. Adaptation tries to reduce the risk that has to be insured. Better drainage can reduce flood damage. Stronger buildings can reduce losses. Better land-use planning can keep expensive infrastructure out of harm’s way. More resilient farming can reduce the financial shock of drought.

The question for India is whether the increasingly detailed information produced by climate and catastrophe models will be used only to charge more accurately for risk, or also to reduce it.

Some risks will remain too large to be carried by an individual insurer, business or government. That is why India is exploring broader ways of sharing disaster losses. The United Nations Development Programme has worked with the National Disaster Management Authority on the design of a National Catastrophe Risk Insurance Pool intended to collate public resources and expand financial protection.

The Finance Commission says disaster insurance outside crop insurance remains at an early stage, held back by limited products, low penetration, lack of awareness and problems of affordability and access. The emerging idea is not that insurers should carry the entire burden. Governments, insurers, reinsurers and other sources of finance may all have to carry different parts of it.

Catastrophe models have a role because risks have to be understood before they can be shared. But the models cannot answer the question of who should bear the cost. That remains a human choice.

Global warming and climate change have ensured that the weather will continue to surprise us. A monsoon may fail in one place and overwhelm another. A drought can damage a harvest while a flood destroys a road. A cyclone can leave thousands of claims in its wake. The purpose of better models should not be merely to become better at sending the bill. It should be to help India build places, farms and businesses that are safer to insure, and ultimately, cheaper to protect.