People tend to assume a life insurance premium is a price, set the way a retailer sets a price: whatever the market will bear. It is closer to an estimate with a margin on top. Every premium is built from two components that behave completely differently, and once you can see the seam between them, a lot of otherwise confusing things about this industry start making sense, including why two policies covering the same person can cost noticeably different amounts.

The short version

  • A premium is mortality cost plus expense load, plus a margin for the risk that the estimate is wrong.
  • Mortality cost is the part that is genuinely about you. It is the least negotiable part.
  • Expense load is distribution, underwriting, administration, and capital. It is where the differences between companies live.
  • A lead sold to several agents raises the cost of every policy written from that batch, because most of the money spent chasing it produced nothing.
  • On an illustration, read the guaranteed column first. The other columns are assumptions.

The two halves of a premium

Start with mortality cost. An insurer looks at a pool of people with broadly similar characteristics and estimates how many claims that pool will produce over time. Divide the expected claims across the pool and you get the raw cost of the promise. This is the part of your premium that is genuinely about risk, and it is the part that is least available for anyone to be clever about. An insurer that underestimates it does not get a bargain; it gets a solvency problem several years later.

Then there is everything else, which the industry calls the expense load. This is the cost of running the business that delivers the promise: finding you, evaluating you, issuing the contract, collecting the money, keeping the records, answering the phone, paying the claim, and holding capital against the possibility that the mortality estimate was too optimistic.

The interesting thing about that split is what it implies. If you want a cheaper policy and you are the same person with the same health history, the mortality half is not going to move much. Everything that can move is in the expense half. Which means that when a company claims it can price better, the honest follow-up question is: which expense did you remove, and how?

Distribution

Finding a person who wants coverage, and compensating everyone involved in the conversation that led to a policy.

Underwriting

Evaluating the case, which includes every case that was evaluated and never became a policy.

Administration

Issuing, billing, servicing, and paying claims for decades, plus the capital held behind all of it.

What distribution actually costs

Distribution is usually the largest single piece of the expense load on a small policy, and the reason is arithmetic rather than greed. A small policy generates a small amount of premium, but the work required to sell it is not proportionally smaller. Somebody still has to find the person, have the conversation, answer the questions, take the application, and follow up. That work costs roughly the same whether the policy is small or large.

There is also a chain, and each link in it takes a share. In a common arrangement, a lead source finds the prospect, an agency or marketing organization recruits and supports the agent, the agent has the conversation, and the insurer pays compensation that flows back up through those layers. None of those parties is unnecessary. Each is doing something. But each one is paid out of the same premium, and the person paying the premium is funding all of them.

This is also why persistency matters so much to an insurer, and why it should matter to you. Most of the distribution cost is incurred at the front, in the first year. A policy that lapses in year three means that cost was spent and never recovered, and the pricing of every future policy has to anticipate a certain amount of that. A book of business that stays in force is genuinely cheaper to run than one that churns, which means a company that sells policies people keep can charge less than one that sells policies people abandon.

What manual underwriting costs

Underwriting has two costs, and the second one is the one people forget.

The first is direct. A human underwriter reading a file is being paid to read it. If the process orders medical records, a paramedical exam, or lab work, those have costs attached, and they are incurred whether the case is approved or declined. Time is a cost too: a case sitting in a queue is a case that might not close, because the applicant loses interest or buys somewhere else in the meantime.

The second is the cost of every case that never became a policy. An insurer does not underwrite only the applications it accepts. It underwrites all of them, and the expense of evaluating the ones it declined has to be recovered from the ones it issued. The higher the ratio of work to issued policies, the more each issued policy has to carry.

This is the mechanism behind automated underwriting

The saving is not that a computer is cheaper per decision than a person, although it generally is. The saving is that a decision made in the conversation removes the queue, removes the paramedical step where the product allows it, and removes most of the cost of evaluating cases that were never going to be issued. Speed is a pleasant side effect. The cost structure is the actual reason it matters, and it only works if the decisions are good, because a model that approves the wrong cases has moved cost from the expense load into the mortality half, where it is far more expensive.

Why a shared lead raises the price of a policy

This is the part of the cost structure that consumers almost never see, and it is worth explaining properly because it is genuinely counterintuitive.

Suppose a marketing company generates an inquiry from somebody interested in coverage, and sells that inquiry to several agents at once. That is a perfectly legal and extremely common business model. From the marketing company's perspective it is efficient: one inquiry, sold several times.

Now look at it from the buyer's side. Several agents call the same person, often within minutes of each other. One of them, at most, writes a policy. The others spent their time and their share of the lead cost and got nothing. But they still have to cover their costs, so the price they are willing to pay for leads has to account for the fact that most of the leads they buy will not convert. That expectation is baked into what the lead costs, which is baked into what the agent needs to earn per policy, which is baked into the compensation structure the product has to support, which is baked into the premium.

In other words: the cost of all the conversations that produced nothing is recovered from the policies that got written. Nobody in the chain is doing anything improper. The structure just means that a person who buys a policy is paying for the several agents who called and did not sell one, plus the aggregator margin on top of all of it.

The same logic explains why an aged lead, resold months later, is cheap and converts poorly, and why an exclusive lead costs more per inquiry and can still be cheaper per issued policy. What matters for a premium is not the price of a lead. It is the total acquisition cost per policy that actually gets written and stays in force.

What that means for you as a buyer

You cannot audit an insurer's expense load from your kitchen table, and you should be suspicious of anyone who suggests you can. But the structure does suggest a few practical things.

  • Cheap is not always cheap. A low first-year premium on a product whose cost rises with age is a different thing from a level premium. Ask which you are looking at.
  • Compare like for like. A level benefit policy and a graded benefit policy are not comparable on price, because they are not the same promise in the first years.
  • A policy you keep is cheaper than a policy you replace. Replacing coverage restarts the front-loaded costs, and may restart contestability. Sometimes it is still right. It is rarely free.
  • Ask what is guaranteed. Not what is projected, not what is current, not what is typical. Guaranteed.
  • Being called by six agents is a signal. It tells you something about where your information went, and about the cost structure behind whatever you are about to be sold.

How to read a policy illustration

An illustration is a table showing how a policy is expected to behave over time. It is the most useful document in a life insurance conversation and the most commonly misread, because it contains two very different kinds of number sitting next to each other in similar-looking columns.

Some columns are guaranteed. They show what the contract obligates the insurer to do in the worst case it is permitted to deliver: the maximum charges it could apply, the minimum interest or dividends it must credit. Those numbers are a floor, and they are the only numbers on the page that are a promise.

Other columns are not guaranteed. They are projections built on current assumptions about charges, interest, or dividends, and those assumptions can change. A non-guaranteed column is not dishonest. It is an estimate produced under rules about how estimates may be presented. It is simply not a commitment, and it is often the column a seller points at.

Six things to do with an illustration

  • Find the guaranteed column and read it first. If the policy does not work for you on the guaranteed basis, you are relying on assumptions.
  • Check whether the premium shown is level for every year of the table, or whether it changes at some point.
  • Look at the death benefit column in the early years, especially year one and year two. This is where a graded structure shows itself.
  • Read the footnotes. The assumptions, the exclusions, and the words "not guaranteed" live there.
  • Confirm the insurer named on the illustration is the insurer on the policy, and that the ages and health class used match you.
  • Ask for a copy to keep, and ask for the policy itself before you sign. The contract governs; the illustration describes.

The honest summary

Most of what a small life insurance policy costs is not the risk. It is the process of finding you, evaluating you, and administering the promise, plus the cost of everyone who was paid along the way and every case that was worked and never issued. That is not a scandal. It is what happens when a product with a small premium has to be sold one conversation at a time through a chain of intermediaries.

It does, however, mean the expense side is where the interesting engineering is. Removing a manual review queue, owning the demand instead of renting it from an aggregator, and declining the cases that do not fit are all ways of reducing cost without reducing the promise. That is the thesis we are building on, and we would rather explain the mechanism than publish a number we cannot yet stand behind.

If you are shopping today, take the framework and use it on whoever is in front of you. Ask what is guaranteed, ask what the benefit is in year one, ask who the insurer is, and ask why the premium is what it is. The answers will tell you a lot about the company as well as the product.