Why Population Longevity Does Not Feel Personal

The IZA experiment turns on a distinction that is easy to erase in ordinary conversation: the lifespan of people similar to us is not the same belief as the age at which we expect to die ourselves. This analysis focuses on that self–other boundary and on the practical question it answers: how can an official population statistic be accurate, credible, and tailored to a respondent’s characteristics without changing the personal belief that feeds a retirement plan? The answer requires separating what the survey observes from the psychological mechanisms the authors propose.

The distinction is not a semantic detail. A population estimate describes an average for a reference group. An own-longevity estimate is a judgment about one person’s place within—or outside—that group. Retirement planning depends on the second judgment, because it is the personal horizon that determines how long someone thinks their assets and pension income must last.

Two questions that look alike but do different work

The survey asked respondents to estimate the average age at death for people with the same relevant age, gender, and county of residence. That question supplied a defined target. Researchers could compare the answer with the official actuarial estimate for the same reference group and measure an error without asking respondents to predict their own death. It was a test of population longevity literacy: could a person form a reasonably accurate belief about the group to which the statistical benchmark applied?

Later, respondents estimated their own expected age at death. That second answer could draw on private information that a population average does not contain, including a person’s perception of their own health and circumstances. It also carries a different psychological meaning. Asking about similar people invites an outward-looking judgment based on observable characteristics. Asking about oneself invites an inward-looking judgment about identity, vulnerability, resilience, and personal exception. The two estimates can be related without being interchangeable.

The control group shows how far apart they can be. Respondents expected to live to about 85 years and four months themselves, around three years longer than the lifespan they assigned to similar people. Twenty-one percent believed that similar others would die before age 80, but only 1% expected to die themselves before reaching that age. The gap is not a technical artifact of a decimal place. People placed a large share of their peers below a threshold that almost none of them applied to themselves.

This pattern is consistent with what the authors call personal optimism: people tend to see themselves as less exposed to adverse outcomes than comparable others. But the important evidence is not only the difference in means. The relationship between the two beliefs was weak. Across untreated respondents, the correlation between own-longevity beliefs and population-longevity beliefs was 0.09. For those who thought similar others would die before 80, it was slightly negative, at −0.03. For those who placed similar others above 80, it was positive but still modest, at 0.22.