Probability turned into a business. It earns twice: pricing risk, and investing the float.
An insurer sells a promise today and usually does not pay on it for years, sometimes decades. In between, it sits on a huge pile of other people's premiums and invests it, money Warren Buffett famously named the float. So an insurer earns in two ways at once: by pricing the risk correctly, charging enough to cover the claims that will eventually come, and by investing the float well while it waits. Picture a giant community pot where the unlucky few are paid by the lucky many, with the pot itself earning a return in the meantime. Because the profit on a policy is spread across its whole life, normal profit and PE mean almost nothing here, so the industry built its own yardsticks: the value baked into each new policy, the embedded value of the whole book, and how many customers keep paying their renewals.
Answer these six and you have understood the industry. The rest of the page is the detail behind them.
An insurer is one of the few businesses where this year's profit tells you almost nothing, and where the company holds a fortune that is not yet its own.
Insurance is, at heart, a community pot. Many people pay in a small premium, and the unlucky few who suffer a loss are paid out of what everyone contributed. The insurer's first job is deciding who is allowed into the pot and what each of them should pay, so that the premiums coming in comfortably cover the claims going out. That job is called underwriting, and getting it right is the whole game.
Here is what makes insurance special. The insurer collects premiums today but pays claims years, sometimes decades, later, so at any moment it is sitting on an enormous pile of other people's money. That pile is called the float, a term Warren Buffett made famous, and the insurer invests it in the meantime, earning a return on cash it has not yet had to pay out. So an insurer makes money twice: once by pricing the risk correctly, and again by investing the float while it waits.
Great insurance products do not sell themselves, and that is the strange thing about insurance. People rarely wake up wanting to buy it. Someone has to reach them and persuade them, at a bank counter, through an agent, or in an app. So the ability to reach millions of buyers, the distribution, matters more here than almost anywhere else, which is why insurers owned by big banks have such an edge: they inherit a ready-made network of branches full of customers. Insurance is sold, not bought.
Because the profit on a policy is spread over decades, you cannot judge a life insurer by this year's profit. So imagine selling one life policy today. You will not know its full profit this year; you will earn it slowly, over the next twenty years of premiums. Rather than wait two decades to find out, the industry estimates that entire future profit today, at the moment of sale. That estimate is called the value of new business, or VNB, and the slice of the premium it represents is the VNB margin. Add up that value across the whole book of policies, not just this year's, and you get the embedded value, which is why life insurers are priced on embedded value rather than PE.
A policy only makes money if the customer keeps paying the renewals. Persistency measures how many stick around. High persistency means the sales were genuine and the book is healthy; low persistency means policies were mis-sold and lapse, wasting the cost of every sale.
Two insurers with identical policies can still make very different money, because of what they do with the float.
It is easy to picture an insurer as just a seller of policies, but the pile of float it invests is often far larger than its own capital, and how well it invests that pile decides a big part of the return to shareholders. Two insurers could price risk equally well and still end up worlds apart, simply because one earns more on its investments than the other. In that sense an insurer is an investment company wearing an insurance uniform: the underwriting fills the pot, and the investing is what quietly compounds it.
Life insurance is long-term: decades of premiums, read on embedded value and new-business margin. General insurance (health, motor, fire) is short-term: you insure for a year and renew, and it is read on the combined ratio, which is claims plus costs as a share of premiums. Below 100% means the insurance itself makes money; above 100% means it loses on insurance and leans on investment income.
The two classic failures are underpricing risk, where a wave of claims wipes out years of premiums, and mis-selling, where policies lapse or the book was built on false promises. Watch the solvency ratio, the cushion that proves the insurer can pay future claims.
For life insurers, use price-to-embedded-value and the new-business margin, plus persistency and solvency. For general insurers, the combined ratio is king. PE will mislead you in both. But step back from the yardsticks and the whole industry fits in three sentences. Every insurer has two jobs: price risk correctly, and invest the float wisely. Fail either one, and the shareholder suffers.
Every sector reduces to a demand metric, a pricing metric, an efficiency metric, a capital metric, and a risk metric. For insurance, these are the ones that matter.
What each number tells you, and how to read it.
| Metric | Why it matters |
|---|---|
| APE Growth | Annualised premium equivalent, the sales-volume topline for life insurers. |
| VNB Margin | Value of new business margin, the profitability of each new policy sold. Above 25% is excellent. |
| Persistency Ratio | Policy renewals. 13th-month above 85% is good. It measures stickiness and the quality of the sale. |
| Embedded Value (EV) | Intrinsic value: adjusted net worth plus the present value of future profits. The PE equivalent for insurers. |
| EV Growth | The long-term compounding indicator. 15%+ EV growth signals a strong franchise. |
| Solvency Ratio | The ability to pay future claims. IRDAI minimum is 150%. Higher means more financial strength. |
Insurance is a business that earns money twice: once by pricing risk correctly, and again by investing the float while it waits to pay claims.