Expected Value for Insurance Decisions

From Premium to Expected Profit

In Fair Games and Break-Even Prices, you compared a player's cost with the expected value of a prize. An insurance policy can be modeled from the insurer's point of view in a similar way: the insurer receives a premium and may pay a claim. The payout is uncertain, so the insurer's profit is a random variable.

For this tutorial, assume the insurer receives one fixed premium for the policy period and has no other revenue or costs in the model. Let \(Y\) be the insurer's payout during that period, and let \(c\) be the premium. The insurer's profit is the premium received minus the payout. A payout of zero means no claim payment; a payout greater than the premium can make the insurer's profit negative for that policy.

The formula follows from weighting each possible profit by its probability. For every outcome, subtract the payout from the same premium. The expected profit is therefore the premium minus the probability-weighted average payout. This connects directly to the expected-value calculations in The Mean of a Discrete Random Variable and to subtracting a fixed cost in Expected Value in Games and Raffles.

The last step uses the fact that the probabilities for all possible payouts add to 1. You can calculate expected profit in either of two ways: find the expected payout and subtract it from the premium, or first calculate the insurer's profit for every payout and then find the mean of that profit distribution. The two methods should agree.

Before calculating, check that every payout probability is between 0 and 1 and that the probabilities for all listed outcomes add to 1, as in Checking Whether a Probability Distribution Is Valid. Also check that \(Y\) means the dollar amount the insurer actually pays, not the customer's repair bill before a deductible or the total value of the insured property.

A Procedure for Modeling Insurer Profit

A probability table may give payouts directly, or it may give claim outcomes that need to be translated into payout amounts. Once \(Y\) and the premium are identified, make the profit distribution by calculating \(c-y\) for each payout \(y\). This makes the perspective of the calculation explicit: money received by the insurer is positive, and money paid out reduces profit.

If the expected profit is positive, the premium exceeds the average payout in this model. If it is zero, the premium equals the expected payout. If it is negative, the average payout exceeds the premium. None of these comparisons says what will happen on a particular policy: even a policy with positive expected profit can produce a large loss when a claim occurs.

Worked Example: Calculate Expected Profit from a Payout Table

Worked Example: Use a Four-Step Solution

Worked Example: Choose a Premium for a Target Expected Profit

What an Expected-Profit Model Leaves Out

A positive expected profit is an average across the modeled chance process, not a guarantee for one customer or one policy. The payout distribution in the examples gives a high probability of no claim and a smaller probability of a substantial payout. That pattern can produce positive expected profit while still allowing individual policies to lose money for the insurer.

The model also treats the premium as the insurer's only income and the listed payout as its only cost. In practice, an insurer may have administrative expenses, taxes, commissions, and other costs. If the model supplies a fixed expense \(a\) per policy, profit could be defined as \(X=c-Y-a\), and expected profit would be \(c-E(Y)-a\). Do not subtract such expenses unless the problem gives them or asks you to include them.

The distribution itself is an assumption about possible payouts and their probabilities. Expected profit is only as useful as that model. A carefully computed answer cannot make an unrealistic payout distribution accurate. In an AP response, identify the model's stated probabilities and interpret the result conditionally: “According to this payout model...” is often a useful phrase.

Common Mistakes and AP Exam Tips

• Using the customer's profit instead of the insurer's. The insurer receives the premium and pays the claim, so its profit is \(c-Y\). Reversing the subtraction gives the other party's perspective.
• Treating the claim amount as the insurer's profit. A payout is money paid out, not money earned. A larger payout reduces the insurer's profit; it can make profit negative.
• Forgetting the zero-payout outcome. If no claim is possible, include payout \(Y=0\) and its probability. Its contribution to expected payout is zero, but the outcome belongs in a complete distribution.
• Subtracting the expected payout from the premium and then subtracting it again. Use either \(c-E(Y)\), or calculate each profit \(c-y\) and weight those profits. Do not combine the methods by subtracting the payout twice.
• Calling expected profit guaranteed profit. A full-credit interpretation gives the unit and the time period, and describes the long-run average rather than one outcome. For example: “The expected profit is $21 per policy year under this model; actual profit on an individual policy may be negative.”
• Claiming that expected profit is the insurer's actual business profit. State what is excluded. If expenses are absent from the given model, say the result is before those expenses rather than silently treating them as zero in reality.

A strong response shows the probability-weighted calculation, makes the insurer's perspective clear, and finishes with an interpretation in context. When possible, calculate expected profit both as premium minus expected payout and as the weighted mean of possible profits. Agreement between the two is a useful arithmetic check.

Key Takeaway

An insurance premium is fixed in this model, while the payout varies according to a probability distribution. Comparing the premium with the expected payout gives the insurer's expected profit, but the result describes an average under the model rather than a guaranteed outcome.