Billion-dollar Rain: Why India can’t afford to ignore urban flood risk

The 2005 Mumbai flood was a watershed event for India that resulted in over 1,000 fatalities and caused extensive damage to infrastructure, businesses and homes. For the Indian insurance industry, it remains the costliest natural catastrophe on record with a total cost of INR 2,250 crore (USD 500mn) in claims at 2005 values. Two decades on, this catastrophic flood continues to serve as a stark reminder of the urban flood risk in the country.

Swiss Re estimates that a repeat of this event today could cost as much as INR 20,000 crore (USD 2.3 bn) – and this is not even the worst-case scenario. Rapid, unchecked urbanisation combined with more frequent short-duration intense rainfall could cause even larger losses, with new hotspots emerge in line with India's growth as a global manufacturing and technology hub.

While the re/insurance sector can help absorb financial losses, building societal resilience to flood risk requires collaboration among multiple stakeholders on urban flood mitigation planning. Developing viable flood insurance solutions requires a holistic and forward-looking risk assessment, using flood catastrophe models and high-quality data that flow across the insurance value chain.

It is important that we reflect on the lessons learned from past events and recognise the growing impact of urban flooding across India. Over the past two decades, floods have accounted for approximately 67% of all economic losses and 66% of all-natural catastrophe-related fatalities in India. Today, individual events that cause economic losses in the ballpark of INR 8,700 crore (USD 1bn) in 2025 prices have become increasingly common.

Geographically, India is naturally prone to flooding due to its varied topography, encompassing plains, hills, coastal areas, and extensive river network. Flooding is often associated with major rivers like the Ganga-Brahmaputra overflowing their banks in the northern and eastern states during the monsoon. However, in recent years, urban centres have witnessed a sharp rise in pluvial flooding, where local drainage systems are overwhelmed by heavy rainfall. Cities like Mumbai and Chennai have already facing serious impacts, and similar scenarios are likely to arise in many more urban centres unless resilience measures are scaled up.

On 26 July 2005, Mumbai experienced an extreme weather event, recording 944 mm of rainfall in a single day – almost 40% of its annual rainfall – coinciding with a high tide and peak hourly rainfall of up to 80mm. With nearly 90% of the city's land surface concretised, permeability was minimal, leaving little room for water to seep into the ground. Mumbai's drainage system, designed to handle 25mm/hr rainfall, proved grossly inadequate for such intensive rain. Additionally, the Mithi river, which flows through populated and industrial areas was severely encroached upon, limiting its capacity to function as a natural drainage channel.

The resulting flood submerged over 30% of Mumbai's area and led to more than 1,000 fatalities along with economic and insured losses of INR 13,500 crore (USD 3 bn) and INR 2,250 crore (USD 500mn), respectively.

Since then, India has experienced a series of severe urban floods – such as the 2006 Surat flood, 2014 Jammu and Kashmir floods, 2015 Chennai floods, 2018 Kerala floods – that have had significant impact on society. These flood events have also led to high insured losses, largely because of the concentration of commercial assets and relatively high insurance penetration in urban centres.

Two key trends are contributing to the increasing frequency and severity of urban flooding in India:

• Unchecked rapid urbanisation: Economic growth has accelerated urbanisation, significantly increasing the concentration of commercial infrastructure in cities. Since 2000, India’s built-up area has expanded by nearly 100%. The compromise of natural drainage systems – such as water bodies and green areas – has increased vulnerability to flood events.
   
• Increasing short-duration intense rainfall: Climate patterns are changing. Scientific studies indicate a notable rise in the number of days with high intensity rainfall globally. A study published in scientific journal Nature Communications suggested a threefold increase in widespread extreme rain events (i.e., daily rainfall >150mm) across central and western India (including industrial states like Maharashtra and Gujarat) during the period from 1950–2015.

Since 2000, India has experienced rapid economic growth, with nominal GDP expanding at an average annual rate of 12%. The property insurance sector has grown in tandem, with premiums increasing at an annual average of 11%. A common method to estimate potential future losses involves trending historical loss figures using economic indicators such as real GDP, inflation, insurance exposure, and premium growth.

Exhibit 3 shows the insured losses for key events if they were to happen in 2025 (i.e. as-if estimates) accounting for the exposure growth using property fire premium changes as a proxy. These estimates, however, do not account for other critical variables such as changes in land use (e.g. increased run-off with growth of built-up area), climate change, or any mitigation and adaptation measures that may have been implemented.

Current expectations suggest that:

• An event with insured losses exceeding INR 2,500 crore (USD ~300mn) may occur every 2-3 years.
• An event with insured losses exceeding INR 5,000 crore (USD ~600mn) may occur every 5 years.
• Insured loss events at the scale of INR 7,500 crore (USD ~900mn) may have a return period of 8-10 years; such events included Jammu and Kashmir (2014; INR 9,000 crore, USD ~1bn), and Chennai (2015; INR 10,500 crore, USD ~1.3 bn).

The 2005 Mumbai flood is the top insured loss event on as-if basis over the last 25 years with losses estimated at INR 20,000 crore (USD 2.3bn), about 8 times the original losses of about 2,250 crore INR (USD 500mn) at 2005 values. 

While the 2005 Mumbai flood remains the top benchmark for insured flood losses, similarly significant or even worse scenarios are plausible. Importantly, severe losses are not solely driven by extreme rainfall totals. Various factors, like reservoir releases during severe flood events; prolonged rainfall over multiple urban centres; unfavorable antecedent conditions: and the interactions of multiple weather systems can contribute to devastating economic and insured losses.

For example, while the 2023 Sikkim flash flood was not included in the analysis due to significant uncertainty in loss estimates, it highlighted potential loss scenarios involving flash flooding in hilly regions with high-valued energy infrastructure. If comprehensively insured, the claims from such an event could rival or even surpass those of the 2005 Mumbai flood.

Based on Swiss Re modelling, we estimate a nationwide insured flood loss of INR 20,000 crore (USD 2.3bn) to have a return period of around 50 years.

Underwriting flood risk is inherently complex. Catastrophe flood models are still evolving to provide a comprehensive view of flood risk. A major challenge is the need for high-quality, location-specific data, which is often lacking—especially in emerging markets like India. Recent flood events, both in India and globally, offer valuable insights for refining underwriting practices in this challenging environment.  

A holistic view of flood risk

Flood risk is not limited to riverine floods, which are typically shown in flood risk maps. It is critical to review other flood types to avoid unpleasant surprises. A comprehensive approach should include urban flooding, cloudbursts, and glacier bursts, particularly in India's urbanised hilly areas. Further, coastal areas face additional threats from storm surges and cyclone-induced flooding. These conditions can be combined with other weather systems to generate unexpected and severe flood scenarios.

Flood risk assessment in a changing environment

Reliance on past events alone is insufficient in assessing flood risk in today's context. Rapid urbanisation continues to reshape India's exposures (as cities like Bangalore, Pune, Chennai, and Hyderabad continually expand), while newer urban centers (like Dehradun, Coimbatore, Guwahati, Kochi) are emerging as hubs; the changing landscape is influencing patterns of flood vulnerability. This means that new flood hotspots may often appear in areas with limited prior risk assessment experience. This dynamic environment requires a forward-looking approach particularly in rapidly growing markets. Advanced catastrophe flood models that accurately reflect present conditions are invaluable tools to address these evolving challenges.

Exposure and loss data in the insurance value chain

Accurate and reliable flood risk assessment is dependent on the quality of exposure and loss data, particularly given the localised nature of flood perils. While data quality in India Insurance industry has improved from state-level to zip code aggregates —a positive step— effective assessment ideally requires street-level or precise latitude or longitude resolution.

Moreover, many basic data attributes such as insurance conditions and coverage breakdowns remain unavailable during modelling. This limits the industry's ability to assess and price risk with confidence.

The 2005 Mumbai flood served as a stark wake-up call for India's insurance industry, marking the most expensive insured flood loss at the time. Since then, unplanned rapid growth and climate change, without corresponding infrastructure, have led to the emergence of new flood hotspots. Similar flood scenarios like Mumbai's and Chennai's experience are likely to occur in many more locations across the country.

Resilient urban planning and robust drainage infrastructure are crucial to reducing future losses and societal risk. The Insurance industry remains a critical partner in mitigating flood risk by offering relevant and innovative solutions. At the core of developing such solutions and fostering a transparent, confident and thriving insurance industry, is access to high-quality exposure and loss data. With the right flow of data and increased use of flood models, the industry can establish continuous feedback loops that drive ongoing improvements in flood risk assessment capabilities, ultimately building a more resilient future.

The information in this article is generic in nature and should not be considered in the form of professional advice. Swiss Re does not guarantee the accurateness and completeness of the information mentioned in this article. Swiss Re shall not be liable for any damages arising out of or in connection with this document. This article is intended for informational purposes only and should not be interpreted as official guidance or advice from SwissRe.