Out in the cold

The recent series of hurricanes and floods have provided some very real demonstrations of climate change. The bad news is that these extreme weather events are not going to go away. Every model that has been produced indicates that we are going to be experiencing more severe weather conditions in the future. One of the major challenges for insurers is to capture sufficient data to make informed decisions on underwriting and rating, in the face of increasingly frequent natural perils.

The good news for insurers is that there have been some significant developments in terms of the data available to them in the past decade. Things have certainly changed since the early 1990s, when many were rating at the level of postal districts for property insurance - which splits the UK into about 2500 areas. Back then, nobody in the market had any sort of data on river flood modelling; these days, there are some insurers that are going as far as rating at individual address level for flood risks. There has also been a massive increase in computing power in the past 10 years, meaning an increase in the complexity of the models and the amount of data they can crunch to produce results.

One of the current problems, however, is that not all insurers are making the best use of the risk databases, and pricing tools available to understand their exposure in greater depth and how climate change might impact them in the future. One of the issues that may hold people back from taking action is that many of the climate models estimate what will happen during the next 25 or even 50 years. There is a sense that if something is not happening tomorrow, or at least within a working lifetime, there is little point in doing much about it. It is important to bear in mind, however, that an incremental increase in gust windspeed by area over the next decade or so could lead to a significant increase in the volume of attritional storm losses.

Equally, it is useful to bear in mind that while it is becoming commonplace to use flood models for rating purposes, it is less common for companies to access information on other natural perils. People have so many things on their 'to do' list that implementing a well-thought-out data strategy is not always top of the agenda but it is going to be crucial. Those insurers that fail to keep an eye on what is happening now and in the future will ultimately lose out.

The information available from risk modelling must go hand in hand with other factors, such as the information gained from talking to insureds, as well as the expertise of the in-house actuary and the underwriter's experience. All this available information can be at odds with a desire to reduce quotation times and to avoid bombarding the insured with too many questions.

The right questions

It is, therefore, important to ask the right questions. If there is no real risk of natural subsidence, or if the property is not on a flood plain or exposed to storms, fewer questions could be asked.

However, if looking at a property built in Harrow in the 1930s, for example, even without any obvious exposure to natural perils, it would be wise to check whether or not there is a tree close to the house. This is because when these properties were built, there was a drive to increase greenery in the area, and a resulting trend to plant trees outside property. Some of these houses now have mature trees just outside their front doors, with the associated subsidence risk this carries. Consequently, the ability to make informed choices about when to ask more detailed questions is key.

In reality, of course, this can be difficult, particularly in a call centre environment - again, it is likely that technology can provide the solution. There are already systems in use that can drive scripts, based on the answers given by the customer. Perhaps in the future this type of technology will be used to flag up problem locations to a call centre operator in a property insurance context, in order to prompt a different line of questioning.

One of the issues that needs to be taken into consideration is how far a property's particular characteristics make it either more resilient or more vulnerable to natural peril risks. For example, one in five properties built on flood plains have been constructed since 1980 - and newer properties tend to be more vulnerable to flood damage. This is because they often have plaster-board internal walls that have little defence against water damage, whereas older houses benefit from solid brick inner walls. Also, about 10% of the properties on flood plains are bungalows, which have high vulnerability to flood due to all rooms being on one level.

As far as windstorms are concerned, the shape of the roof is a vital factor in determining potential damage. The age of the property is also significant - the older the property, the more likely the materials used to build it have deteriorated over time.

Home improvement

A particular consideration today, in light of the trend for home improvement, is that a property's vulnerability can increase if modifications are made to the building. A simple conservatory extension can greatly increase the chance of a windstorm claim. Asking the relevant questions to ensure data is up to date has to be part of every renewal.

We are still learning about the implications of climate change but one thing seems certain - weather conditions are not going to stabilise. It is important for the insurance industry to be receptive to new risk-modelling methods and to be aware of the questions it must ask its customers. It is also key that the data captured, whether from third parties or directly from insureds, is as accurate as possible. There is an old maxim that should certainly be applied - if you put garbage in, you will get garbage out.

Mark Harrison is managing director of ISL, an Experian company.