One of the main issues underlying insurance contracts is moral hazard: insurers have no control over policyholders' behaviour, but the likelihood of paying a claim very often depends on the latter. The reverse is also true as individual behaviour is affected by insurance coverage: if people are insured, their exposure to danger could increase because they have fewer incentives to try to prevent accidents from happening (Stiglitz, 1983). Moral hazard is striking evidence that a solution (i.e. insurance coverage) exacerbates the problem (i.e. risk exposure) that the solution was supposed to solve.
A case in point is motor insurance. What drivers do while steering is completely beyond the control of the insurance company, and the risk of claims increases when drivers know that they can ask for financial compensation in the event of an accident. Traditionally, companies have tried to mitigate moral hazard through rewards and penalisations, such as no-claims bonus systems and deductibles. In both cases, the basic idea is to encourage policyholders to keep their exposure to dangers under control and to retain, at least in part, responsibility for what happens.
Digital technologies promise to transform the way insurance companies deal with moral hazard. In third-party liability motor insurance, this has already been happening for several years through the use of telematics. According to many scholars, telematics data-generating devices should enable insurance companies to technically control moral hazard and improve their claims prevention capacity. The role of insurers is therefore going to be increasingly proactive: instead of stepping in ex post to compensate for damage, insurers can intervene ex ante by warning policyholders of the dangers they are exposed to and, at the same time, offering advice and safe-driving solutions to improve individual driving style.
Insurance companies offering telematics motor insurance policies term this type of service "coaching". What they mean by this is a twofold strategy. Firstly, telematics devices are used to provide real-time driver coaching. In 2021, for example, Ford launched SYNC 4®, a voice assistant installed on E-Transit vehicles. It alerts drivers with beeping alarms when they exceed a speed limit or fail to fasten a seat belt. Insurers, in turn, are increasingly focusing on advanced driver assistance systems (ADAS) because the telematics data these systems produce can be processed to calculate a risk score, which, being a risk assessment tool, could improve actuarial calculations and complement existing rating methods.
A further strategy is represented by telematics data promoting safe(r) behaviour over a longer period of time. Through an application installed on a mobile phone, information extracted from telematics data is fed back to the driver after every trip. The application not only translates individual driving behaviour into a score, but also shows possible criticalities (e.g. harsh braking, dangerous tailgating, sudden lane-changing, exceeding speed limits etc.), thus conveying information that is expected to help drivers improve their driving skills and mitigate reckless attitudes. The application acts as a kind of instructor that corrects drivers when they make a mistake or encourages them when they drive carefully. While it was once thought that many road accidents were due to drivers' miscalculations (of driving skill, road conditions or individual ability to control the vehicle), telematics data now conveys the feeling that it is possible to calculate drivers' miscalculations and use the results of this calculation not only to monitor individuals when they are behind the wheel, but also to somehow change their behaviour.
.jpg)
If we look more closely at telemetry package-based motor insurance, however, things look different. Telematics does not change much in terms of control over the individual driving behaviour. Insurers are not able to intervene like a driving instructor when coaching a new driver. Rather, what insurers can do is try to control their lack of control over policyholder behaviour. This can be achieved in two ways: (1) through the feedback the policyholder receives from the insurance company via an app. In this case, information produced by telematics data is used to enhance the policyholder's self-control. However, recent studies show that feedback alone is not enough; therefore, (2) insurers can combine information with financial incentives, such as a discount upon policy renewal or some form of cashback (Stevenson et al., 2021). Intrinsic motivation (i.e. avoiding breaking your neck) is too weak if it is not reinforced by some form of extrinsic motivation (i.e. saving a few dollars).
Even if changes in the behaviour of individual drivers with telematics insurance policies were detectable in the long run, it would be difficult, if not impossible, to say with certainty whether this change was the result of coaching (i.e. an effect of the information), the desire to save money or simply chance. Long-term studies and sound statistical evidence would be needed to draw such conclusions. Like behaviour, a customer's motivation remains beyond the control of insurers.
Telematics represents a unique opportunity to improve risk assessment, foster safer behaviours and reduce the information asymmetry underlying the relationship between insurers and policyholders. Above all, insurance companies will have to increase research on the predictive value of telematics data and find the best way of communicating this prediction to policyholders. They should not forget, however, to take into consideration the possible effects that communicating the results of predictive analytics could have on policyholders' behaviour and their relationship with insurers.
Alberto Cevolini is adjunct associate professor at the University of Bologna. He was fellow of the Alexander von Humboldt Foundation and visiting professor at Bielefeld University. He is responsible for the research project on "Personalised Insurance", which is part of the main research project PREDICT on "The Future of Prediction: The Social Consequences of Algorithmic Forecast in Insurance, Medicine and Policing" funded by an ERC Advanced Grant (ERC-2018-ADG, n. 833749, P.I. Prof Elena Esposito). More information on Prof. Alberto Cevolini can be found here.
References
Stiglitz, Joseph (1983): "Risk, Incentives and Insurance: The Pure Theory of Moral Hazard". The Geneva Papers on Risk and Insurance, 8(26), pp. 4–33.
https://link.springer.com/content/pdf/10.1057/gpp.1983.2.pdf
Stevenson, Mark et al. (2021): "The Effects of Telematic Based Feedback and Financial Incentives on Driving Behaviour: A Randomised Trial". Accident Analysis and Prevention, 159, pp. 1–8.
https://www.sciencedirect.com/science/article/abs/pii/S0001457521003092?via%3Dihub
