Risk Transfer Basics
Insurance is a bad deal on average. It has to be — the insurer pays every claim, salary and profit out of your premiums. You buy it anyway for the risks where 'on average' is the wrong way to think.
Why you buy something with a negative expected value
Every insurer takes in more in premiums than it pays out in claims. It has to — claims, staff, marketing and profit all come out of the same pot. So on average, across everyone who buys a policy, buying it loses money.
That makes insurance irrational for any loss you could simply absorb. If a $400 repair would be irritating but survivable, paying someone a margin to cover it is just paying the margin. This is the whole case against extended warranties, phone insurance and very low deductibles.
It flips completely when the loss would end you. A house fire or a liability judgment is not a bad month, it is the end of your financial life — and there is no average to fall back on, because you only get one draw. Paying a small certain cost to remove a small chance of ruin is a good trade even though the expected value is negative.
Not 'will this happen?' but 'if it did, could I absorb it?' Insure the losses you could not.
A deductible is the part you kept
The deductible is the slice of every loss you have agreed to carry yourself. Raising it hands more risk back to you, and the insurer discounts the premium because it now expects to pay less.
That trade is arithmetic, not preference. If raising a deductible by $2,000 saves $400 a year, you are being paid $400 annually to accept $2,000 of one-off exposure. Five claim-free years and you are ahead; a claim in year one and you are behind.
The catch is that it only works if the $2,000 exists in cash. A high deductible funded by a credit card is not risk retained, it is risk ignored — you took on the exposure and gave up the point of the savings. This is why the emergency fund comes before the deductible decision, not after it.
Insure what would ruin you; self-insure what would merely annoy you. The deciding question is never how likely a loss is, but whether you could absorb it if it happened.
Sources
- Consumer Insurance Resources — National Association of Insurance Commissioners (2026)