Moral hazard – it’s one of those theoretical concepts that you learn at some point in an intermediate microeconomics course, but that gets increasingly fuzzy in your memory as time passes. The authors of a really nice recent paper on the history of moral hazard make the point that the use of the term actually varies greatly across professional and academic disciplines. In economics, moral hazard is the financial risk to the insurer that arises from (rational) agents doing less to avoid losses when insurance is present. It is often contrasted with adverse selection, which is the financial risk to the insurer arising from loss-prone types of agent buying more insurance than their more fortunate peers. Moral hazard is about the incentives under insurance, not about any characteristics of the insured themselves.
The paper shows that moral hazard is used differently in the insurance literature, and how the ‘moral’ component has come to capture any kind of behaviour that can be interpreted as taking advantage of the insurance provider. For moral hazard in the economics definition, the canonical example is torching an insured building; for adverse selection, the canonical example is a risk-loving or unhealthy person buying life insurance. I think these examples clearly show how the term moral hazard can be used as a powerful economics buzzword to express indignation at immoral behaviour. But where do these moral values come from? Well, the paper also points out that insurance companies have incentives to take part in the public debate that shapes our attitudes towards those people that ‘take advantage’ of insurers. In this debate, the threat of higher insurance premiums is a powerful argument. Remind you of anything?
I think that the economics distinction between moral hazard and adverse selection is theoretically convenient (it sharpens your thinking) but the practice can be quite fuzzy (the paper makes this point, too). In empirical applications, it is hard to find out which of the two effects is at play (although there is a literature on identification strategies). Consider a car rental company that introduces new insurance allowing for greater excess reduction and subsequently finds that damage is more common for customers who buy the extra cover. If you don’t know anything about the drivers’ abilities, it’s hard to say whether the extra accidents are caused by average drivers being more careless or bad drivers buying more insurance. From the point of the rental company, there is a simple solution – the insurance premium for the extra excess reduction will be adjusted to match the damage frequency (insurance that matches expected values is called actuarially fair). The alternative action – screening customers on their safety record – is used by some companies to exclude drivers but is not used (as far as I know) to set the price of insurance.
Anyway, I digress. The paper is well worth a read.
