The SOA Longevity Illustrator vs. a Health-Adjusted Estimate: What Each Tells You

Two Good Tools That Answer Slightly Different Questions

If you want a quick, credible read on how long you might live, the Society of Actuaries Longevity Illustrator is one of the best free tools available. It is built by actuaries, it is grounded in real mortality data, and it gives you something far more useful than a single number. It deserves its reputation.

It is also not the same thing as a health-adjusted life expectancy estimate, and the difference is worth understanding before you lean on either one. The Longevity Illustrator asks a handful of questions and returns a population-based answer tilted by a general sense of your health. A health-adjusted model asks many more questions about your specific conditions and builds an estimate around them. Both are legitimate. They simply answer slightly different questions, and they are useful at different moments.

This article explains what the SOA Longevity Illustrator does well, where its coarse health input runs out of resolution, and when it makes sense to reach for a more granular, condition-level estimate instead.

What Is the Society of Actuaries Longevity Illustrator?

The Longevity Illustrator is a free online tool published jointly by the American Academy of Actuaries and the Society of Actuaries. You can find it at actuaries.org/longevity-illustrator, and it takes about a minute to use. It asks for your current age, your sex, whether you smoke, and a general rating of your health on a short scale — roughly poor, average, or excellent. For couples, it asks the same questions about each partner.

What it returns is genuinely thoughtful. Instead of a single life-expectancy figure, it reports the probability of living to a range of ages, a set of planning ages tied to specific survival probabilities, and, for couples, the odds that at least one partner is still living at each age. Under the hood, it is built on Social Security Administration mortality data with future mortality improvement applied, so it is looking forward rather than freezing today's death rates in place.

In other words, the actuaries who built it made several of the right choices. It is probabilistic, it is forward-looking, and it is honest about uncertainty. For a free tool that takes a minute, that is a great deal of thoughtfulness.

What the Longevity Illustrator Gets Right

It would be easy, in a piece written by a company that builds a more granular model, to damn a free tool with faint praise. That is not the intent here. The Longevity Illustrator does several things genuinely well, and any fair comparison has to start there.

It is free and actuarially grounded. There is no account, no paywall, and no marketing funnel. The methodology comes from two of the most credible actuarial bodies in the country, and it rests on recognized mortality data rather than ad hoc assumptions. As a first stop for a longevity gut check, it is hard to beat.

It thinks in percentiles, not a single number. This is the tool's best feature and the one most people overlook. A single "your life expectancy is 86" figure invites the reader to plan for exactly that age, which is a mistake, because life expectancy is the center of a wide distribution rather than a due date. By reporting the probability of reaching 80, 85, 90, and beyond, the Longevity Illustrator nudges users toward the same distributional thinking that Monte Carlo simulation formalizes. Planning against a percentile is far safer than planning against a mean.

It handles couples well. Joint-and-survivor longevity is one of the hardest things for people to reason about, because the probability that at least one of two people is still alive is much higher than either individual probability. The Longevity Illustrator models this directly, which makes it a useful tool for retirement conversations where the real planning horizon is the second death, not the first.

It looks forward, not just at today's mortality. Because it applies mortality improvement rather than relying on a static period table, it avoids one of the most common errors in longevity estimation, which is assuming that death rates never fall. That alone makes it more realistic than reading a raw Social Security period table, which is still how many people estimate longevity. For more on why those raw tables mislead, see why SSA tables fall short.

If all you need is a fast, credible sense of the range of plausible lifespans, especially for a couple, the Longevity Illustrator is an excellent choice, and this article is not trying to talk you out of using it.

Where a Single Health Rating Runs Out of Resolution

The Longevity Illustrator's design is also its limitation. To stay simple enough to use in a minute, it compresses the entire question of your health into one coarse input: a general self-rating of poor, average, or excellent. That single category has to stand in for everything a full medical history contains.

Consider two 68-year-old men who both describe their health as "average." One has well-controlled high blood pressure, takes a statin, walks daily, and has two parents who lived into their nineties. The other has type 2 diabetes with early kidney involvement, a prior heart attack, and a body-mass index in the obese range. To the Longevity Illustrator, these two men are effectively the same person. They enter identical inputs and receive identical curves. In reality, their true longevity profiles can differ by a decade or more.

The point is not that self-rated health is meaningless. Research consistently shows that how people rate their own health does predict mortality to a degree. The point is that it is coarse. It cannot distinguish between a manageable chronic condition and a life-limiting one, it cannot see the difference between one condition and four, and it cannot capture the way conditions interact. A tool with one health input is doing the best it can with one health input. It is simply working at a lower resolution than some decisions require.

What a Health-Adjusted Estimate Adds

A health-adjusted life expectancy estimate is built to work at that higher resolution. Rather than compressing health into a single rating, it models the individual's actual profile in detail, and that detail changes the answer in ways a coarse input cannot.

Condition-level detail and severity

A health-adjusted model evaluates specific diagnoses rather than a general impression, and it grades each one by severity. Well-controlled hypertension and congestive heart failure are not the same input, and neither are early-stage and metastatic cancer. The same diagnosis can move an estimate modestly or materially depending on how advanced it is, and comprehensive condition modeling captures that difference instead of averaging it away.

Comorbidity interactions

Human mortality is not a linear sum of independent conditions. Diabetes and cardiovascular disease amplify each other through shared biological pathways; several conditions together often imply a worse outcome than adding their individual effects would suggest. A model that reasons about these interactions produces materially different, and generally more realistic, estimates for the multi-morbid clients who are most common in the age groups advisors serve. A single health rating has no way to represent "these three conditions together."

An insured-lives foundation

A health-adjusted estimate can start from a mortality table built specifically for individual life underwriting rather than for the whole population. The SOA 2015 Valuation Basic Table, combined with the MP-2021 mortality improvement scale, provides a granular, forward-looking baseline broken out by age, sex, smoking status, and underwriting class. That is a different and more individualized starting point than population data, though it too is only a foundation on which the condition-level adjustments are built.

A full distribution, not a point

Like the Longevity Illustrator, a good health-adjusted model refuses to reduce longevity to one number, but it goes further, running a Monte Carlo simulation to produce a full survival distribution with explicit confidence intervals and survival probabilities to each age. The result is a personalized version of the same percentile thinking the Illustrator encourages, centered on the individual's actual risk profile rather than on a population tilted by one health category.

Longevity Illustrator vs. Health-Adjusted Estimate: A Fair Side-by-Side

Put plainly, the two approaches differ along three axes.

The first is the input surface. The Longevity Illustrator uses a small, fixed set of inputs: age, sex, smoking, and one health category. A health-adjusted model uses dozens, spanning specific conditions, their severities, lifestyle factors, family history, and functional status. More inputs mean more setup, but also more resolution.

The second is the population basis. The Longevity Illustrator adjusts population mortality data by a health tilt. A health-adjusted model starts from insured-lives actuarial tables and adjusts for the individual. Neither basis is wrong, but they describe different reference classes, and for an affluent, individually healthy person the population basis and the insured basis can point to meaningfully different numbers.

The third is what you do with the result. The Longevity Illustrator is designed for a quick, credible read that anyone can generate in a minute. A health-adjusted estimate is designed to be a planning input you can defend: the denominator in a withdrawal calculation, the horizon in a Social Security claiming decision, the mortality assumption behind an insurance or settlement analysis. The extra resolution earns its keep precisely when a real dollar decision rides on the number.

So Which Should You Use?

Both, at different moments. For a fast, credible sense of the range, the kind of thing you might do at the start of a planning conversation or to help a couple reason about joint longevity, the Society of Actuaries Longevity Illustrator is a fine and honest choice. It will give you percentiles, it will not overstate its certainty, and it costs nothing.

When the number stops being a curiosity and starts driving a decision, the coarse health input becomes the binding constraint, and it is worth moving to a condition-level estimate. Withdrawal-rate math divides the portfolio by a longevity figure. Social Security claiming turns on where a person falls relative to a breakeven age. Life insurance and settlement decisions depend on the expected timing of the death benefit. In each of those cases, the difference between "average health" and a specific profile of diagnoses and severities can change the recommendation, and a tool that cannot see that difference cannot help you make it.

A Note on the Population Basis

One quieter distinction deserves a mention, because it affects who each tool serves best. The Longevity Illustrator rests on Social Security mortality data, which reflects the entire population. A health-adjusted model can start from insured-lives experience, which reflects people who passed medical underwriting and therefore skews healthier. For clients drawn from the upper end of the income and wealth distribution, which describes most advisory clients, the population basis tends to understate longevity before health is even considered. That is the same reason why SSA tables fall short for affluent, healthy individuals, and it is worth keeping in mind when a population-based tool and a health-adjusted one disagree.

Try Both

The best way to understand the difference is to run your own numbers through each. Generate a read on the Society of Actuaries Longevity Illustrator for the percentile view, then try the free longevity calculator to see how condition-level, health-adjusted modeling shifts the same person's distribution. If a real decision is riding on the answer, a full longevity report adds confidence intervals and planning horizons, and advisors can request access to bring health-adjusted longevity into every client plan.