The price already assumed a story about you
Every premium you have priced so far in this rung quietly leaned on a mortality table — a statement about how likely the insured is to die in each future year. But *whose* mortality? A lifelong smoker with a heart condition and a marathon runner who has never touched tobacco do not share the same future. If you charge them the same price, one of them is being robbed and the other is getting a gift. Underwriting is the process of finding out enough about an applicant to decide which story the price should assume.
Recall the engine that makes insurance work at all: risk pooling. Many people each pay a modest premium, and the unlucky few who suffer a loss are paid from the common fund. That magic only holds if everyone in the pool is paying a premium that honestly reflects the risk they bring into it. Drop a person whose true risk is five times the pool's average into a pool priced for the average, charging them the average price, and they are silently subsidised by everyone else. Underwriting is the discipline that guards the gate so the pool stays fair to the people already inside it.
Preferred, standard, substandard — and decline
The output of underwriting is a risk class. The middle bucket is standard: the ordinary, expected-mortality applicant the base table was built for. Below it in risk — and so below it in price — sit the preferred classes, reserved for applicants whose health, habits, and family history are visibly better than average: a non-smoker with clean blood work, normal blood pressure, and no worrying family history. Above standard in risk are the substandard classes, for applicants who carry extra hazard — a chronic illness, a dangerous occupation, a risky hobby — and who therefore pay a higher premium for the same cover. And beyond all of these is the rare outcome of *decline*: an applicant whose risk the insurer simply will not take at any ordinary price. This whole machinery is risk classification.
How does a substandard premium actually get set? The two most common conventions are worth knowing because they reveal how actuaries think about extra mortality. Under the multiple-of-standard (or table-rating) method, the underwriter judges the applicant's mortality to be, say, 150% of standard, and the substandard rate is built from a mortality table scaled up by that factor. Under the flat-extra method, a fixed dollar amount per thousand of cover is added on top of the standard premium, which suits a hazard that is roughly constant year to year — a dangerous hobby, for instance — rather than one that worsens with age. The choice is not cosmetic: it changes how the surcharge behaves over the life of the policy.
Same person, same cover, three risk classes (illustrative annual rate per $1,000)
Preferred non-smoker $0.90 <- best health, habits, family history
Standard $1.30 <- the base table was built for this life
Substandard, 200% table $2.60 <- twice standard mortality assumed
Substandard, flat-extra $1.30 + $3.00 = $4.30 <- +$3 per $1,000 for a
constant hazard (e.g. hobby)
For $250,000 of cover, the standard annual premium is
250 units x $1.30 = $325.
The 200% substandard life pays 250 x $2.60 = $650 for the very same payout.
The gap is not a penalty -- it is the honest cost of that life's risk to the pool.Classification fights adverse selection
There is a deeper reason underwriting is not optional, and it is one you have already met: adverse selection. The applicant always knows more about their own health than the insurer does. People who privately suspect they are high-risk are precisely the ones most eager to buy generous cover, and at the standard price they are getting a bargain. If the insurer cannot tell them apart from the genuinely healthy, the high-risk crowd in, claims run higher than the price assumed, the premium must rise, the healthy then leave because the price no longer suits them — and the pool spirals. Underwriting is the information-gathering that breaks this spiral before it starts.
This also explains a pattern you saw earlier in the survival-models rung: select-and-ultimate mortality. A life that has *just been underwritten and accepted* dies less often, for the first few years, than an otherwise-identical life who was underwritten long ago — because the recent medical screening has, for a while, filtered out the people who were already sick. Underwriting does not merely sort people; it temporarily *lowers* the mortality of those it just admitted. That select period is the statistical shadow that classification casts onto the mortality table itself, and a careful pricing actuary blends select rates, not ultimate ones, into the early years of a freshly issued policy.
Be honest, though, about what classification is and is not. It is *not* a moral judgement and it is *not* perfect prediction. A preferred class is not a club of virtuous people; it is a group whose *average* mortality is lower, and plenty of individuals inside it will outlive plenty inside the substandard class. Classification works on populations, not prophecies. And there are firm limits society places on it: most jurisdictions forbid classifying on race, and many restrict the use of genetic test results — because fairness *to the pool* and fairness *to the person* are not always the same thing, and the law draws the line.
The contestability period: a window, not a trapdoor
Underwriting rests on the applicant telling the truth. But what if they don't? The law's answer is the contestability period — typically the first two years after a policy is issued. During this window, if the insured dies, the insurer may re-open the application and check it for material misrepresentation: an undisclosed cancer diagnosis, a hidden smoking habit, a falsified age. If a lie is found that would have changed the underwriting decision, the claim can be denied or the policy rescinded. This is the contestability period, and it is the legal teeth behind honest underwriting.
Here is the part people get wrong, so hold it carefully. After the two years are up, the policy becomes *incontestable*: the insurer can no longer void it for a misstatement on the original application, even one later discovered to be false. That sounds like a loophole for liars, but it is the opposite — it is a deliberate protection for honest policyholders. It means a family decades later, grieving and filing a claim, will not have the payout snatched away over some immaterial slip on a form filled out a lifetime ago. The period balances two fairnesses: long enough to deter and catch fraud, short enough that an honest contract eventually becomes truly unbreakable.
Group life: underwriting the crowd, not the person
Everything so far has assumed *individual* underwriting — one applicant, one medical exam, one carefully chosen risk class. Now meet the great exception: group life insurance, the cover your employer provides to every staff member, often with no medical questions at all. How can an insurer skip the underwriting it just spent four sections defending? The trick is that the *group itself* does the underwriting for free. People do not join a company in order to get life cover; they join to work. That incidental, non-insurance reason for the group to exist is what tames adverse selection.
Three structural features do the work that an individual medical exam would otherwise do. First, *automatic, near-universal participation*: when cover is employer-paid and everyone is in, the sick cannot selectively pile in while the healthy stay out — the pool is the workforce, not the worried. Second, an *actively-at-work requirement*: an employee must be well enough to be on the job on the day cover starts, a light filter that quietly excludes those already gravely ill. Third, *limits and underwriting only on the excess*: cover up to a generous guaranteed-issue amount is automatic, and only an employee who wants *more* than that ceiling must answer medical questions. The insurer underwrites the group's composition — its age mix, industry, and size — rather than each life.
The trade-offs are honest and instructive. Group cover is cheap and blessedly easy to obtain — a real gift to people whose health would make individual cover costly or impossible. But it is usually *term* cover tied to your job: leave the employer and it typically lapses, though a conversion right may let you buy an individual policy without fresh underwriting. And because there is little individual screening, group cover does not show the same deep select-period dip that individually underwritten business does. The lesson generalises beautifully: whenever a non-insurance reason forces a whole crowd into a pool together, you can underwrite the crowd and skip the individual — which is exactly the principle on which all social insurance rests.
Why this quietly shapes every price
Pull the threads together and a clean picture emerges. The premium formulas of this rung take a mortality table as a *given*; underwriting and classification are how that given is *earned*. They decide which table applies to you, they lower it temporarily through the select period, and they protect its honesty through the contestability window. Loosen the underwriting and adverse selection seeps in, claims drift above the assumed table, and the price you carefully built starts to leak — not in year one, but slowly, the way an underwriting mistake always reveals itself: late, and only in the data.
There is a tension worth carrying forward, too. Sharper classification — more classes, finer distinctions — makes each price more *individually* fair, but it also shrinks each pool and chips at the very risk-spreading that makes insurance worthwhile; pushed to the extreme, everyone pays their own exact expected cost and the pooling vanishes. Looser classification spreads risk more broadly but invites the healthy to feel overcharged and walk. Every product in this rung sits somewhere on that dial. Where it sits — how hard it underwrites, how many classes it offers — is one of the quiet design choices that, alongside dividends, riders, and lapse, ends up shaping the price on every policy people actually buy.