The Fundamental Insurance Equation

Turning the question around

Everything you have learned so far in the non-life rungs has asked one question: *how much will the claims be?* Frequency and severity, the loss ratio, the loss-development triangle — all of it stares backward at losses that have already happened or forward at losses still to come. Ratemaking turns the question around. It does not ask what the losses *were*; it asks what a policy *should cost next year* so that the price is fair to the customer and survivable for the insurer. That flip — from measuring losses to setting a price — is the whole craft of pricing, and it begins with a single equation so simple you could write it on a napkin.

The equation is just an accounting identity dressed up as a promise. Over a large enough book of business, across a future policy period, the premium an insurer collects must be enough to pay all the claims it expects, cover the cost of running the company, and leave a modest, planned margin for profit and for the cost of the capital that backs the promise. Written out, that is the [[fundamental-insurance-equation|fundamental insurance equation]]: *Premium = Losses + Expenses + Profit*. Nothing in this whole field contradicts it; everything in ratemaking is an elaboration of how to fill in those three boxes honestly and in advance.

The pure premium: the price of the risk itself

Of the three boxes, losses come first, because the price of a risk begins with the risk. The [[pure-premium|pure premium]] is the part of the premium meant to pay claims and nothing else — the expected loss for one unit of exposure, before a single penny is added for expenses or profit. It is the rawest possible answer to *what is this risk worth?* Strip away the salaries, the commissions, the margin, and what is left is the bare cost of the promise: the average amount you expect to pay out per policy.

Here is where the frequency-severity split you already know pays off beautifully. The pure premium is simply expected frequency times expected severity: how often a policy has a claim, multiplied by how much the average claim costs. Suppose, across a book of auto policies, a car has a claim about 8% of the time in a year, and the average claim costs 5,000. Then the pure premium is 0.08 times 5,000, or 400 per car-year. That single number is the expected cost of one car's risk for one year — the irreducible core that the rest of the price is built around.

You will also hear the term [[loss-cost|loss cost]], and in everyday pricing work it means almost the same thing: the expected loss per unit of exposure. The shade of difference is one of convention. A loss cost often bundles in loss adjustment expense — the cost of settling claims — because that cost rides along with the loss itself; and in many markets an advisory organisation publishes *loss costs* for an entire industry, which each insurer then marks up with its own expenses and profit to reach a final rate. So 'pure premium' tends to be the textbook term for expected loss per exposure, and 'loss cost' the working term for the same idea as it travels through filings and rate manuals — frequently with LAE already folded in.

From pure premium to the price on the bill

The pure premium pays the claims, but it cannot pay the people who sell the policy, the staff who run the company, or the taxes the government levies — and it leaves nothing over to reward the capital that stands behind every promise. So the actuary grosses it up. Expenses come in two flavours that matter for the arithmetic: *fixed* expenses, a flat dollar amount per policy regardless of price (a policy-issuance cost, say), and *variable* expenses, which scale with the premium itself (commissions are a percentage of premium; so are premium taxes). On top of both sits the [[underwriting-profit-provision|underwriting profit provision]] — the planned margin, also usually a percentage of premium, that lets the price honour the *Profit* box of the equation.

Because variable expenses and profit are percentages *of the final premium itself*, you cannot just add them on — they would themselves need topping up. The clean trick is to divide. Add up every cost that scales with premium — variable expenses plus profit — call that the *variable cost ratio*, and the share of premium left to cover the pure premium and the fixed expenses is one minus that ratio. This share has a name you will lean on constantly: the [[permissible-loss-ratio|permissible loss ratio]], the fraction of each premium dollar the actuary *permits* to be spent on losses, the target the price is built to hit.

Gross premium from a pure premium (one car-year)

   pure premium (loss + LAE)        =  400
   fixed expense per policy         =   25
   variable expenses (commission,
      tax) = 18% of premium
   profit provision                 =   5% of premium

   variable cost ratio  = 18% + 5%   =  23%
   permissible loss ratio = 1 - 0.23 =  0.77

   gross premium = (pure premium + fixed) / permissible loss ratio
                 = (400 + 25) / 0.77
                 = 425 / 0.77
                 = 551.95  per car-year

Look at what fell out: a pure premium of 400 became a charged price of about 552. The extra 152 is not the insurer being greedy — it is the cost of running the operation and a slim, deliberate reward for putting capital at risk against a year that might go badly wrong. And notice the elegance of the permissible loss ratio: it closes the loop straight back to the loss ratio you already know. If reality matches the assumptions, claims-and-LAE should come in at 77% of premium, the variable costs and profit make up the other 23%, and the profit margin materialises exactly as planned. The pricing target and the after-the-fact scorecard are two faces of the same coin.

The three goals a rate must balance

A rate is not just a number an actuary likes; in almost every jurisdiction it is a number the law constrains. Insurance regulators the world over set the same three standards, often in nearly identical words: a rate must be adequate, not excessive, and not unfairly discriminatory. These are the legal embodiment of fairness, and they pull against one another, which is exactly why pricing is a balancing act and not a calculation. Together they are the heart of rate adequacy and equity.

1. Adequate — the rate must be high enough to cover the expected losses, the expenses, and the cost of capital. An inadequate rate is not generous; it is dangerous. It quietly drains the insurer's surplus and, in the worst case, leaves real claimants unpaid when the company cannot meet its promises. Solvency, not stinginess, is what the adequacy standard protects.
2. Not excessive — the rate must not be unreasonably high relative to the risk. An insurer cannot use market power, customer inertia, or a captive market to overcharge far beyond what the losses, expenses, and a fair profit justify. This standard protects the policyholder's wallet and is the direct counterweight to adequacy: push the price up forever and you are always 'adequate', but you become excessive.
3. Not unfairly discriminatory — policyholders with the same expected cost must pay the same price, and those with different expected costs may, and indeed should, pay different prices. The word 'discriminatory' is technical, not pejorative: charging a young, high-frequency driver more than a low-risk one is *fair* discrimination grounded in cost. What is forbidden is charging two genuinely identical risks differently, or building the price on a variable the law deems off-limits. This is the equity that classification ratemaking exists to deliver.

Hold the three together and you see why no single number satisfies everyone. Adequacy pushes the price up; the not-excessive standard pushes it down; the not-unfairly-discriminatory standard governs how the total is *split* among policyholders rather than its overall level. The fundamental equation gives you the size of the pie — how much premium the whole book must raise. The third standard is about cutting the pie fairly: each policyholder should pay in proportion to the risk they bring. A good rate honours all three at once, which is far harder than it sounds and is the reason a whole profession, not a spreadsheet, sets prices.

What this equation does not tell you

The fundamental equation is true and indispensable, but it is honest only if you remember its silences. First, every term in it is an *expectation* about a future that has not happened. The pure premium is an average; the actual losses of any one policy, or any one year, will scatter wildly around it — that is the whole reason insurance exists. The equation balances *in expectation, over a large book*; it never promises that this year's premium will exactly equal this year's losses plus costs. A bad year can blow straight through an adequate rate, and that gap is precisely what surplus and reinsurance are there to absorb.

Second, the equation says nothing about *time*. Premium is collected up front, but claims are paid later — sometimes years later, especially in long-tailed liability lines. Money has a time value, so a rigorous rate discounts future losses back to today and gives the insurer credit for the investment income earned on funds it holds in the meantime. Two of the rungs you already climbed — interest theory and the loss-reserving work — are what let you fill the *Losses* box at its proper present value rather than its raw, undiscounted face amount.