Quality of reinsurers’ data ‘worse than thought’
Poor data is holding the industry back—but the industry should not accept that this should be the case, Jason Futers, CEO of Insurdata, told Monte Carlo Today.
“When Insurdata started in mid-2017, we knew the industry had a big problem with the quality and relevance of exposure data being used in its core activities. However, it’s actually worse than we realised,” said Futers.
He said this opinion is based on experience. For one client, a major industry player, Insurdata analysed its regional portfolio and found that over 90 percent of its exposures were not geographically where they were believed to be when the policies were bound—with an average displacement error of 54 metres per exposure.
This particular portfolio was in Florida, a region where, Futers believes, the frequency and severity of natural catastrophes should incentivise the industry to understand the quality of its exposure information better—and to collect as many relevant exposure attributes as possible.
“We’ve analysed data in 80+ countries for multiple perils and we see a similar picture everywhere,” he said.
One of the ways poor data holds the industry back is the disruption caused when an event occurs.
“For example, the HIM hurricanes created months of chaos as companies battled legacy systems and other challenges to identify the required data. They were asking: where are my risks? What damage was caused? How can I mitigate risk going forward?” he said.
“This process is very expensive and disruptive—costing time, money and focus that could be directed at internal challenges such as growing a profitable portfolio, or societal issues such as the protection gap. Having quality exposure data should be considered one of the fundamental targets for re/insurers.”
Technologies that Futers believes are conducive to better underwriting include parametric technologies and Machine Vision.
“One of the advantages of Machine Vision, for example, is how the technology can be used to effectively support the experience and knowledge of the underwriter,” he said.
“Supporting the existing knowledge is key to new technology adoption in insurance. Machine Vision can significantly increase efficiency within the underwriting process.”
For better portfolio management, Futers said, blockchain is one technology that can help solve the challenge of the collection and transfer of exposure data, which is a very labour-intensive process that is repeated every year.
“Exposure data can be stored, independently verified and published to a blockchain, therefore facilitating the very efficient capture and transfer of data to anyone accessing that blockchain,” he said.
“For portfolio management, if you consider some companies have tens of millions of risks globally, this efficiency has a huge benefit, freeing experienced analysts from the manual work of translating and manipulating spreadsheets.”
Since its inception, Insurdata has created its technology stack and data architecture to embrace a range of technologies.
“Parametric policies require a range of technology solutions, including hardware with the likes of flood gauges and drones, and technology such as Machine Vision. Our objective is to create key exposure attributes for all lines, all perils, globally. Machine Vision is one of the key components of our technology agenda,” Futers added.
“Our technology is blockchain-enabled, which allows us to confidently build our tech components without fear that we would one day have to rebuild to include blockchain technologies, and to be able to publish our data to whichever blockchain is adopted by the industry.”
