Buying back the tail

Back in track 4 you sat on the other side of this table. You set premiums, paid claims, watched the underwriting result sit stubbornly at a small loss, and still finished the decade richer because of what the float earned while it waited.

Everything you learned there still holds. The only thing that changes now is which chair you are in.

Which means the buyer is knowingly taking the worse side

This follows from the arithmetic and it surprises people, so it is worth saying plainly.

If the premium exceeds the expected claim, then averaged over everyone who buys the policy, buyers pay more than they get back. Insurance has a negative expected value for the customer, by construction, always.

Predict

Given that, what is the correct conclusion about buying insurance?

Choose one answer

It is a bad deal and should generally be avoidedIt is a bad deal on average and can still be the right purchase, because it removes outcomes you could not come back fromIt is only a bad deal if the insurer is overchargingIt depends on whether you are likely to claim

Lock in answer

What follows: transfer severity, carry frequency

Once you know the average is against you, the rule almost writes itself. Every unit of risk you transfer costs you the margin and the expenses. So transfer only the units where the alternative is unacceptable, and carry the rest.

Transfer the tail. Events that would end the position: the house, the liability claim, the long illness, the death of an earner with dependants. Rare, severe, and impossible to self-fund.

Carry the middle. Events that would be irritating: the phone, the appliance, the small excess. Frequent, cheap, and self-funded out of a reserve. Every policy covering these charges you a margin for handling money you could have handled yourself.

The excess is the dial between the two. Raising it hands the frequent, cheap part of the risk back to yourself and keeps the rare, expensive part with the insurer, which is the shape you wanted.

What would you do

A household is renewing cover on a thing worth about 25,000. Three quotes, differing only in the excess — the amount they pay themselves before the insurer pays anything.

Excess 250 costs 620 a year. Excess 1,000 costs 430 a year. Excess 3,000 costs 290 a year. They have made roughly one claim in the last decade.

Optimising for

Lowest total cost over a decade, subject to never facing a bill you could not pay from a reserve of 6,000.

You cannot know

• How many claims the next decade actually brings
• Whether claiming affects what the next renewal costs
• Whether the reserve will still be intact on the day a claim happens
• How the insurer's own claim estimate compares to the household's

Choose an option

Excess 250, at 620 a year.Excess 1,000, at 430 a year.Excess 3,000, at 290 a year.No cover, and save the premium.

The four risks and the one budget

The harder version of the decision is not the excess. It is choosing which risks get transferred at all, when the money available to spend on premiums is finite.

What would you do

A household with two children, one earner bringing in 55,000, savings of about 20,000, and a home with a large loan against it.

Four risks are uncovered, and quotes exist for all of them: cover for the earner dying, cover for the earner being unable to work for a long period, cover for the home burning down, and cover for the contents.

Optimising for

Removing the outcomes that would end the household's financial position, with 1,500 a year of premium available.

You cannot know

• Which of the four is most likely for this particular household
• Whether the second adult could return to full-time work, and at what pay
• How the loan against the home would be treated if the income stopped
• How long the savings would actually last under each of the four events

Choose an option

Cover the contents and the home first — they are the visible assets.Cover the loss of the earner and the long-term inability to work, then the building, and carry the contents.Cover everything at the cheapest available level for each.Cover nothing and add the 1,500 a year to savings instead.

›Two things the insurer knows that you should price inOptional

Adverse selection. The people most eager to buy a specific cover are disproportionately the ones who expect to claim on it. Insurers price for this, which means the honest, low-risk buyer subsidises it. On a policy where self-selection is strong, the premium is worse relative to your own risk than the headline suggests.

Claim friction. The expected claim in the pricing is what the insurer expects to pay, not what customers expect to receive. Exclusions, evidence requirements and the sheer effort of claiming all sit between the two. Cover that is difficult to claim on is cheaper for a reason, and the reason is not generosity.

Worth remembering

• A premium is the expected claim plus expenses plus margin, less what the insurer earns on the float while it holds the money.
• Insurance therefore has a negative expected value for the buyer by construction, and can still be the correct purchase.
• What is being bought is not an average — it is the deletion of outcomes there is no coming back from.
• The rule that falls out: transfer severity, carry frequency, and use the excess as the dial between them.
• Thin cover on a catastrophic risk is the worst combination — a premium every year and an unpayable number afterwards.