Driving Down the Price of Risk
Companies that insure our road vehicles request information including the driver’s age, gender (no longer legal in Europe), claims history and the ZIP or post code where the vehicle is parked at night. On this narrow data set, insurers construct an analytic model used to assess and price risk. A ruling from the European Court of Justice thinned an already sparse model: from 21 December 2012, insurers in that region can no longer vary price of their premiums based on gender.
The immediate effect has been an increase in premiums charged to women in the UK by as much as £500 as reported by the Guardian newspaper. The existing model also disadvantages the young who, as new drivers, haven’t yet had the opportunity to establish records as safe road users. Using age as its determinant, the model groups potentially low-risk drivers and reckless drivers together, and prices their risk similarly. Undoubtedly, some young people drive dangerously – and neurological research explains why. Adolescents have reasoning ability close to that of adults, but they display high rates of poor decision-making: “in risky driving scenario, teens increased risk when tested in company of friends and adults didn’t.”
The current model used for vehicle insurance is imprecise: reliant on a small data set, it is ill-equipped to identify low-risk individuals within a cohort of higher-risk drivers. Further, the recent EJC ruling exposed the model’s fragility: women driving on Europe’s roads on 22 December 2012 represented no greater risk than they did on the previous day, but because the insurers’ pricing tool captured so little useful information, it responded to a legal change by abruptly increasing prices. Models capturing larger data sets would demonstrate greater flexibility.
Launched in May 2010, insurethebox pioneers a data-driven model of vehicle insurance in the UK market. The company fits telematics technology to measure how, where and when each insured vehicle is driven, and this information is delivered to a warehouse for analysis. The company’s data scientists assess drivers individually and make evidence-based decisions to price their insurance. It transpires that driving behaviours such as fast acceleration and deceleration, and taking corners at speed are better indicators of a high-risk driver than age, gender or post code.
Insurethebox further differentiate their product by feeding back results of their analyses to the data creators – their customers – via personalized, secure web portals. The company advises customers what actions they can take to reduce their risk of accident – and the cost of their insurance premium – and promotes good driving behaviours to cultivate. While useful to drivers of all ages, this information may be particularly valuable to younger drivers, who according to Professor BJ Casey, Director of Sackler Institute for Developmental Psychobiology, suggests respond better when their good decisions are rewarded rather than bad ones are punished.
Insurethebox’s use of big data and analytics to derive a better price for risk is working as a business model. Based on a survey of 1498 of its customers undertaken between June and November 2012, the company found they save on average £601 on their car insurance. According to data from the UK Government , the median gross weekly earnings for 18-21 year olds in full-time employment is £280 per week or £14,560 a year – telemetry-based insurance has the potential to save these young people more than 4% of their pre-tax income.
While privacy concerns deter some consumers from adopting telemetry-based insurance, economics makes it too attractive for young people to ignore. Feeling which way the wind blows, other companies are following insurethebox’s pioneering lead. I think we are experiencing only the beginning of a wave – data from existing in-vehicle systems will swell the telemetry stream, showing, for example, whether a driver engaged their vehicle’s indicator light signals before changing lane, then braking hard to avoid collision. Growing data volumes and deeper analyses will create further opportunity to refine the pricing of risk and coach for better driving.
In the vehicle insurance sector the older model used to price risk is limited by knowing too little. Understanding an individual’s driving behavior informs precise pricing of risk – analysing larger data sets reduces the price of vehicle insurance. Big data is creating opportunities across industries; pioneering companies will create new business models, disrupt established markets, and create value for customers and shareholders.
To learn more about how big data can help insurers, download Harness the Power of Big Data for Insurance