In this post I will argue that most consumer-level insurance is a waste of money. If insurance companies are able to make a profit, then the expected value of an insurance policy to the consumer must be negative. As long as we are not at risk of bankruptcy, we should cut out the middleman and just foot the bill in the event of a loss.

Company cars

Some years ago, I was working at a large multi-national company. For trips off-site, rather than pay to hire a car from an external firm, one would book a car from the internal company pool. (This is another example of cutting out the middleman.) The cars were not insured with an external insurer. If a worker were to crash the car, the company had enough money set aside to simply replace it. Hedging this risk would cost money, and is not required for a company with sufficient assets.

The situation is different for individuals. A particularly unfortunate car accident could result in liabilities of order millions of pounds, enough to bankrupt almost anyone. A consumer needs to hedge the risk of motoring to eliminate the possibility of being wiped out. The same applies to other potentially catastrophic losses such as house fires, claims of professional negligence and, apparently, injury to celebrity body parts.

Policies make prizes

Most game shows take out insurance policies to cover top prizes. This makes sense as a worthwhile hedge: TV budgets are fixed, but hedging allows the show to offer a higher top prize to attract more viewers. For games of chance the probabilities involved are known by both parties, so it must be quite difficult for the insurer to make a profit. On the other hand, for games of skill or knowledge, probabilities are approximated in a Bayesian approach where priors are subject to debate.

Your weakness, exploited

The most exploitative kind of insurance is that covering breakage of consumer products like phones and domestic appliances, or hired equipment like skis and bicycles. Often the policy is pushed onto consumers using the pernicious practice of upselling. To a naive buyer, the cost seems small relative to the purchase price and therefore a good deal — but the seller would not be offering the insurance if it didn’t make them more money.

Consumer product insurance takes advantage of our human tendency for risk-aversion. We attach a fictitious mental cost to the uncertainty of possibly having to pay for a new phone in the event that it breaks, and are willing to pay insurance premiums to placate our irrational brains. This is despite the fact that we could afford to replace the phone without an insurance payout.

Beating the dealer

By pooling risk from many policy holders, insurance companies have plenty of data and the law of large numbers on their side. It is therefore difficult to (legally) outwit an insurer. However, I shall discuss two possible exceptions.

Risk-aversion is not the only cognitive bias that insurers can exploit: inertia is another. It is a standard tactic to steadily raise insurance premiums after each yearly renewal, not to reflect greater risks, but because insurers know that switching is a hassle that many people will choose to avoid. The premium increase is then free money for the insurer.

A corollary of policyholder inertia is that discount deals may be had for new customers. Given consistent price rises for loyal customers, it is possible that they subsidise new customers to the extent that, at least for the first year, a policy is worth it. One would have to switch annually to take advantage of this.

The second exception is where the policyholder can make a better estimate of risk than the insurer. Giving the insurer incorrect information to this end is of course fraud; I have one legal but contrived example. On average men, especially young men, are demonstrably more dangerous drivers than women. It does not seem unreasonable to me that men should pay more for car insurance, but modifying pricing on gender grounds is illegal under equality law. Instead, insurers look for other characteristics that are closely correlated with gender.

Suppose your occupation is nursing, and you drive a pink car. The insurer will look at the average motorist with those attributes, and infer that you are a careful driver. If you happen to also be male, then you will be undercharged relative to your true statistical risk.

It seems unlikely that either of these two effects are large enough to make insurance yield a positive expected value, but they at least reduce the cost of a hedge you might need to take anyway.

The house edge

I conclude with the assertion that taking out an insurance policy is functionally the same as gambling on sports or playing the Lottery. However, the entertainment value of insurance is much lower, and so is even less financially sound! Many sensible people I know consider betting a reckless activity, but are happy to slowly bleed money on insurance premiums for their phone. With insurance, you stake money on the fact that a loss will be incurred, with rather long odds set by the insurance company. The event of a payout for a successful claim leaves you better off compared to being uninsured; the more likely outcome is that your premiums are pure profit for the insurer. At a casino, the house always wins in the long run.