Flood alerts: Using real-time data to mitigate the cost of flood insurance

We live in a wet country. Thanks to our temperate climate, we get cool, wet winters and warm, yet still wet, summers.

As all house insurers know, the consequence of this is that flooding occurs, and cannot be avoided. According to a 2023 report by the UK Health Security Agency ^[1], over six million people currently live in areas which are at risk.

These numbers are set to increase dramatically as more houses are built on flood plains and, more particularly, as we experience the impact of global warming. According to the same report, the number of people in the UK significantly at risk of flooding will rise by 61% by 2050 under a modest warming scenario (2°C) and an even more dramatic 118% in a high warming scenario (4°C).

The need to improve flood risk profiles

It is clear that this will have significant implications for insurers. Being able to improve flood risk profiles – not just for individual applications at the point of sale but across their entire portfolio – is going to become increasingly critical to their operations.

Currently, insurers rely on stochastic event sets (SES) which enable them to create statistical models of the hydrological dynamics of a particular catchment area. In other words, they can use theoretical modelling techniques to determine the allocation of flooding risk.

But theory, as they say, is one thing; real-world practice is often another thing entirely.

Real-world events simply don’t show up in the stochastic models. And besides, flooding events happen very rapidly. Ideally, insurers should be able to know about them in real-time to make an immediate decision on how to price for the risk of that property. They should also be able to use that real-time, and recent historical insight at the point of quote to mitigate their expose to risk.

It’s true that the various national environmental agencies issue regular flood alerts and, moreover, they update them at 15 minute intervals. But as anyone familiar with the BBC weather forecast will know, these tend to cover huge areas. Although more granular, individual post code alerts are available, but they are only offered to individuals making a one-off enquiry via the relevant agency website.

Experian Flood Alerts

Insurers need a different solution. They need to be able to access the entire alert data from the agencies and, in real-time, associate it with their portfolio as a whole, including each and every individual address that they are potentially selling insurance to right now.

This is the gap that the new Experian Flood Alerts product fills. When a flood occurs, the Experian Flood Alerts system issues the appropriate flag alerts to insurers at the individual property level, giving them critical real-time and recent history context.

This delivers several critical benefits to insurers:

Protecting insurers at the point of quote

The real-time insights provide an extra layer of protection to insurers at the point of sale. Underwriting engines can automatically call our Application Program Interface (API) to check on the current and recent real-life status of an address. In fact, every 15 minutes, insurers can update their decline codes at the property level to reflect changing conditions on the – potentially increasingly sodden – ground. Because this can be done at the property level it is a far more granular, and therefore accurate, view than post codes data can offer. It’s also worth noting that, without this real-time data, insurers are typically vulnerable to customers exploiting the 14 day cooling-off period by taking out speculative insurance. With Flood Alert information, however, the theoretical risk of stochastic modelling can be immediately updated based on current and recent flooding events.

Helping insurers prepare for the ‘demand surge’

When insurers can see that a significant amount of their portfolio is at risk of flooding they need to respond to the predicted ‘demand surge’ for goods and services. Here, the key objective for insurers is to reduce cost. That means doing whatever is required to speed up the clean-up operation. After all, few things are as expensive as rented accommodation and hotels for those whose homes are uninhabitable. It also means acting swiftly to reserve resources such as dehumidifiers and tradespeople. The sooner insurers get the data they need to act the better, as the cost of these must-have resources can sky-rocket tenfold following a major flood.

Continuous improvement: using real data to refine theoretical models

Statistical models might indicate an area is liable to flood once every 100 years. Experian’s 10-year enrichment data, however, may indicate that the same area – perhaps because of new buildings on a flood plain – has received a flood warning every year for the past decade. In other words, the real-time updating can also act as a crucial monitoring/verification system enabling insurers to dynamically alter their risk profiles to more accurately predict where flood may occur.