Moral Hazard

Study guides: Exam 6C, Exam CP-321, Exam GH-101

Moral Hazard is the change in an insured's behaviour caused by being insured: less care taken to prevent a loss (ex ante), or more cost incurred once a loss has occurred (ex post), because someone else pays. It is a behavioural consequence of the contract, not a selection effect at purchase — that is Adverse Selection.

Ex ante moral hazard — reduced loss prevention. An insured property owner may not install the sprinkler; a fully insured driver may drive less carefully. Usually modest in personal lines, because the insured still bears injury, inconvenience and deductible.
Ex post moral hazard — inflated cost after the loss. The larger effect in practice: extended treatment under accident benefits, repair rather than the cheaper alternative, longer disability duration. It requires no dishonesty at all, only the absence of a reason to economise.
Provider-side moral hazard is often the dominant version: the medical, legal, towing and repair providers respond to an insured payer, not the claimant. Much of what is described as claimant behaviour in auto insurance is provider behaviour.
Controls: deductibles and coinsurance (the insured retains part of every loss), experience rating and no-claims discounts (today's claim raises tomorrow's price), policy limits, exclusions, treatment protocols such as the Minor Injury Guideline, and claims investigation.
Distinguish from fraud. Moral hazard is a rational response to incentives and is legal; fraud is misrepresentation and is not. Programs are evaluated on how well they manage the former, not on how much of the latter they catch.
Moral hazard is the standard criticism of government programs whose benefits are generous and whose funding is not experience-rated — the argument against Employment Insurance regional benefit variation and against uniform Workers Compensation Insurance assessment rates, and the reason disaster assistance is said to discourage flood mitigation.

Deductible as a Moral Hazard Control An insurer observes that policyholders with a $\$500$ deductible have a claim frequency of $0.082$ and those with a $\$2{,}500$ deductible have $0.049$ . Average claim severity (ground-up) is $\$4{,}200$ and $\$5{,}800$ respectively.

How much of the difference is moral hazard, and how much is something else?

Answer Very little of the raw difference is necessarily moral hazard — and identifying the other causes is the point of the question.

Truncation, not behaviour. A $\$2{,}500$ deductible eliminates every loss below $\$2{,}500$ from the reported count. Losses that occur but are not reported are not prevented losses. This alone explains lower frequency and higher observed ground-up severity, since only larger losses survive the filter.

Adverse Selection in reverse — self-selection. Policyholders who choose a high deductible are typically those who believe they are good risks, and who are better able to absorb a small loss. The groups differ before any behaviour changes.

Genuine moral hazard, being the residual: with $\$2{,}500$ at stake, the insured takes more care and does not submit marginal claims.

To isolate (3), the analyst must compare losses above $\$2{,}500$ only in both groups, and control for the risk characteristics that drove the deductible choice. If frequency above $\$2{,}500$ is still lower in the high-deductible group after controlling for those, the remainder is behaviour.

The general lesson for evaluating any insurance program: an observed difference in claim experience between two coverage designs is mostly truncation and selection, and the behavioural effect is what remains after both are removed.