Some companies offering autonomous vehicle solutions may reach a size whereby they will benefit from cost-saving advantages if they create their own captive to manage the business, as opposed to insuring it in the commercial insurance market.
This was the view of the panel discussion entitled, ‘The Cognitive Captive: Artificial Intelligence for Smarter Insurance’, which took place at the VCIA annual conference in Burlington, Vermont.
Speakers on the panel included: Tracy Hassett, president and CEO of Educators Health; Stan Smith, leader at Milliman’s gradient AI, predictive modelling practice in Boston; and Peter Tomopoulos, senior manager at Deloitte Consulting’s actuarial, rewards and analytics practice.
“While captives generally do not replace existing commercial tens insurance, captives fill an unmet need in the marketplace,” said Tomopoulos.
The panel noted that autonomous vehicle manufacturers and commercial fleet operators may reach a scale that allows them to self-insure, similar to what large transportation and logistics companies do today. It noted that there is an increasing incentive to create and use captives and other risk pooling mechanisms.
Tomopoulos noted that companies may want to pool their risks into a captive or RRG given the many players involved, for example the car manufacturer, original equipment manufacturers, car dealers and vehicle owners.
He added that if an accident occurs, it may be costly to discover and litigate at fault, due to multiple sensors in a vehicle from multiple OEMs.
Captives will lower the cost of insurance and reduce risk of the parent’s balance sheet, particularly for first dollar coverage.
“Insurers will want a premium for first dollar or low deductible coverage given the uncertainty involved,” said Tomopoulos. He added that the risk a highly specialised and uncertain, and insurers may not have the data from the manufacturers to be able to charge a fair price.
Looking forward, he suggested insurance products will transform as subscription and usage-based models become more prevalent. And captives in particular have more flexibility to tailor policies and ability to bundle products.
The panel also highlight the prevalence of predictive analytics in the insurance industry, something for captives to consider. 82 percent of carrier respondents currently use predictive analytics 49 percent within personal auto, 37 percent within homeowners, and 32 percent within commercial property/casualty, according to the ISO 2013 Insurance Predictive Modelling Survey.
The same survey showed that 100 percent of big insurers (over $1 billion in premium) are using predictive analytics, and 69 percent below $1 billion in premiums are using it. Predictive analytics were mostly used for pricing and loss cost modelling, followed by underwriting, marketing, claims, and reserving.
VCIA, Vermont, Captives, Ride sharing, Artificial intelligence, Technology, North AMerica