Could technology eliminate moral hazard in car insurance?

A couple of weeks ago, Bloomberg reported that car insurers are offering discounts to insured drivers who agree to have the equivalent of aircraft black boxes installed in their cars:

Smartphone applications and devices that record trip and vehicle data are set to infiltrate auto insurance at a rapid pace, bolstered by discounts of as much as 30 percent. Consultancy Oliver Wyman forecasts that car insurance using driver data to set prices will grow 40 percent a year to become a $3.6 billion market by 2020.

Why would car insurers do this? You can be sure it isn't out of the goodness of their hearts, so there must be something in it for them.

One thing that the insurers are trying to do is to overcome moral hazard - the tendency for someone who is imperfectly monitored to take advantage of the terms of a contract (a problem of post-contractual opportunism). Drivers who are uninsured have a large financial incentive to drive carefully and avoid accidents, because if they have an accident they must cover the full repair cost themselves (not to mention the risk to life and limb). Once a car in insured, the driver has less financial incentive to drive carefully because they have transferred part or all of the financial cost of any accident onto the insurer (though the risk of injury remains, of course). The insurance contract creates a problem of moral hazard - the driver's behaviour could change after the contract is signed.

Now, car insurers aren't stupid and insurance markets have developed in order to reduce moral hazard problems. This is why we have excesses (deductibles) and no-claims bonuses - paying an excess or losing a no-claims bonus puts some of the financial burden of any accident back on the driver and increases the financial incentive for driving safely. This is also why driving illegally usually voids an insurance policy.

However, despite these contract 'enhancements' moral hazard remains a problem for car insurers. The problem remains because the insured drivers' driving behaviour isn't able to be perfectly monitored by the insurance company - they don't know if you're driving safely or not (that is, the asymmetric information about your driving behaviour remains).

This is where new technology comes in. If a black box is installed and the insurance company has ready access to the collected data, then there is little information asymmetry remaining as drivers won't be able to hide their misbehaviour from the insurance company. Now, the black box doesn't let the insurance company know who is driving the car, but since the insurance company is really insuring the car and not the driver it matters little since they should price the insurance policy on the way the car is driven. If your cars turns out to be consistently driven in a risky manner, then you can expect a higher insurance premium to compensate the insurance company for the higher risk. So, the moral hazard problem will be reduced (but not eliminated - there is still an incentive for insurance fraud, and now a new incentive for tampering with the black box).

What's to stop the risky drivers from simply opting out of having a black box? That way, the insurance company wouldn't be able to tell they are driving unsafely, right? Wrong. Since the black box comes along with a premium discount for those who install it, low-risk drivers have an incentive to have the black box installed - they needn't be worried that the insurance company will find out that they are low risk (but they might be worried about the security of their driving data being held by insurance companies!). High-risk drivers want to avoid the insurance company knowing they are high risk, so are less likely to agree to having the black box installed. So, the low-risk and high-risk drivers sort themselves in a way that is advantageous to the insurance company - it helps the insurance company overcome the adverse selection problem.

The adverse selection problem arises in car insurance because the uninformed party (the insurer) cannot tell those with 'good' attributes (low-risk drivers) from those with 'bad' attributes (high-risk drivers). To minimise the risk to themselves, it makes sense for the insurer to assume that everyone is a high-risk driver, and price their premiums accordingly. This leads to a pooling equilibrium - low-risk and high-risk drivers are grouped together because they can't easily differentiate themselves. However, the black boxes solve this problem by causing the low-risk and high-risk drivers to separate themselves in terms of who agrees to have a black box installed (a separating equilibrium) - the low-risk drivers will choose to install the black box, while the high-risk drivers will not.

The insurance companies have also chosen an interesting way of framing this option for consumers. They could have described it as higher premiums for high-risk drivers, but instead they frame it as a discount for those who install the black box. On the surface, this makes it sound a lot more attractive to consumers, since a 30% discount for installing the black box probably seems a whole lot better than a 43% penalty for not installing the black box (even though they are mathematically equivalent). However, it would be interesting to see how drivers would respond to framing this the other way (a penalty for not installing the black box). We know that people are loss averse, and more willing to avoid losses than they are willing to receive equivalent gains - so framing it as avoiding a penalty might actually encourage more consumers to install the black box, as consumers try to avoid the penalty. On the other hand, insurance companies prefer to insure low-risk drivers, so attracting new insurance contracts with low-risk drivers by enticing them with a discount is probably a better move overall. Either way though, moral hazard is likely to be reduced.