Carriers caught out by lack of exposure data

“Carriers—in fact the entire insurance vertical—have been caught off-guard by the need for high-resolution, relevant exposure data,” says Jason Futers, chief executive officer of Insurdata.

Speaking to FIDES Today, Futers warns that there’s been so much focus on modelling within analytics, that some of the basics have received insufficient attention.

This lack of focus has been exacerbated by the ongoing changes in the industry, with natural catastrophes, mergers and acquisitions, and regulatory developments requiring considerable effort, he says.

Futers adds: “Exposure information is simply not fit for purpose; quality is reduced as it transfers through the industry, with considerable human intervention, and when the data arrives at its destination, the geocoding information is inaccurate or insufficient, and necessary attributes, such as first-floor elevation, are simply not available.”

According to the executive, analytics have increased in accuracy and relevance, including accumulation analysis and modelling, over the last 20 years.

As models have matured, the volatility within models has reduced and the resolution has increased significantly, he says, adding that as property can now be analysed and modelled at 2-metre resolution for certain perils, the need for accurate exposure information is paramount.

“There has been very little focus on exposure information until now and the current challenges of exposure information are very similar to what they were 20 years ago. Very little has changed,” he adds.

Futers goes on to warn that the effect of this inaccurate data is “pervasive”, affecting all key metrics in vital areas of pricing and transferring risk.

Insurdata’s analysis to date has shown that 20 percent of geocodes are incorrect by 25 metres or more and, when analysing exposures for accumulation zones and modelling loss estimates, the impact is huge.

He adds: “Missing out on profitable business because a particular accumulation is incorrectly deemed to be full is a significant misuse of capital. Likewise, incorrectly identifying exposures closer to (or further away from) a peril source, such as the coastline, can be the difference between having a profitable year or not.”

There are also considerable operational inefficiencies connected with the management of exposure information.

“We’ve worked with highly skilled, well-paid employees who spend a considerable percentage of their time manually manipulating exposure information. This is a widespread activity within teams from underwriting to exposure management and cat modelling,” Futers explains.

He adds that from Insurdata’s work in this area, its clients have seen a reallocation to higher-value tasks of 50 percent of underlying cost.

Building a rich understanding
Stakeholders across the re/insurance value chain are becoming increasingly aware of the need to use and store accurate exposure information, facilitated by technology.

In five years, companies should be in a position to have built a rich understanding of their key exposures, claims Futers.

“This understanding will build over time, starting from the fundamentals of geospatial location and developing to include key information for all attributes. In five years, this will no longer be the exception but rather the rule, and the data will increasingly be used for risks globally in all the key areas of underwriting and transferring risk,” he predicts.

Focusing on data quality in Latin America, Futers admits that it may not be as high as in some other regions. However, he adds, data resolution is a global problem and while the US is correctly said to have the highest quality data, it’s still “very poor”.

There are a number of reasons for Latin America lagging behind other regions, including the availability and resolution of models across the region.

“Compared to where natural and manmade catastrophe models have been available for many decades, the focus on Latin America is more recent. That has changed and we will continue to see new perils being modelled in Latin America at a resolution that will encourage the use of high-resolution exposure data,” he concludes.