Quantifying driver’s behavior further to better determine bespoke premium for motor insurance
(From memory, this experiment was conducted in 2017/2018. The experiment results was not very promising, as the data collected was too raw to be effectively used to derive insights into driving behavior)
This experiment aims to explore if it is possible to derive further depth of quantified data from telemetry devices embedded or installed onto a car, with the goal of using this data to better determine where the driver’s behavior falls in the spectrum conservative seal and foolhardy cowboy. With this further depth of data, insurers can potentially offer more competitive motor insurance premium to the conservative seal, while using economic incentive to nudge the foolhardy cowboy to adopt safer driving behavior.
Currently, the underwriting process for motor insurance is similar to life insurance, whereby the insurer have historical claims rate of a country’s population, and and further putting them into different strata of risk levels according to varied features. The varied features can be, but not limited to (1) age group; (2) gender orientation; (3) alcohol consumption; (4) history of claims. Each insurer will tweak the features and its threshold to arrive at its own bespoke algorithm for underwriting.
Armed with the algorithm, the next step is to request the potential insured to declare these information. These information can then be fed into the algorithm to determine the risk level, and the associated insurance premium.
The above process had provided a reasonable balanced risk sharing between insurer and insured, despite is a broad stroke approach. However, part of the reason why the broad stroke approach was used is due to excessive difficulty to gather further quantitative data into a person’s driving behavior in order to perform more granular level of risk classification, as it will require a means to observe the driver’s behavior while driving.
Advancement in technology have the potential to address this difficulties, in particular with telemetry devices that can be embedded or installed onto a car. This telemetry device can record further information about a car’s movement, such as average and peak driving speed, velocity at making turns. These telemetry data can be combined with historical information about road condition at the same time to determine how a driver behave under varied traffic conditions.
Such telemetry devices, while appear to be intrusive, is already commonly used in transportation industry. Some are simply counter will increment should the driver exceeds certain speed limit, some recording travel distance and fuel usage to determine fuel efficiency, and some even recording if a driver had veered away from planned route. Hence, the technology is already well tested, and have the potential to be replicated to personal automobile.
For this project, an insurance company is partnering with a ride hailing company to explore if the telemetry device embedded in modern day smart phone can be used to record necessary information about vehicular movement. They are partnering with a FinTech with expertise to interpret these telemetry data into behavioral insights of drivers.
If the experiment is successful, it can provide a further means for insurers to offer more bespoke risk premium tailored to the driving behavior of a insured driver. This can provide more competitive premium for insured drivers that display the characteristics of a conservative seal, while nudging insured drivers that display characteristics of foolhardy cowboys to adopt more conservative driving habits through economic discipline and tailored behavioral advice.
