Are predictions of death greatly exaggerated?

Anew study crosses the transom at Cosh HQ: it is from Britain’s Institute for Fiscal Studies, a microeconomic think-tank. But stop a minute before you turn to the horoscope, or to someone else’s bloviating. This is research with interesting universal implications, research about the most cosmically engrossing topic of all: DEATH.

Specifically, Cormac O’Dea and David Sturrock wanted to know how good people are at anticipating it. And the U.K.’s Office for National Statistics happens to perform long-term surveys that allowed them to find out.

That sounds like a mighty big job for an Institute of Fiscal Studies, but “subjective expectations of survival,” as O’Dea and Sturrock put it in the title, are an important economic subject. In general people have to plan to make their fortunes last at least the rest of their lives, and if they have no clue what those words mean, that is potentially a big social problem.

Moreover, in Britain, as in Canada and elsewhere, the state is giving individuals more freedom in when they begin collecting public pensions. Depending on how long a recipient actually lives, that decision is a sort of bet on, or against, oneself. Most people do not bother to adjust the date they begin to cash in, even though our knowledge of our own health and circumstances should leave us in a superior betting position. You might be missing a chance to rip off the government! Can a columnist write a more profound sentence than that?

In any case the findings of the IFS study are interesting in their own right. And in some ways they are encouraging. Individuals aged 50-plus in the English Longitudinal Study of Ageing (ELSA) — a multi-wave study with a base of about 16,000 respondents — were asked to estimate their personal chances, expressed as a number between 0 and 100, of surviving to particular ages. Everyone was asked about an age marker that was between 11 and 15 years ahead, and the younger respondents were also asked about their chances of having an 85th birthday. Individuals were quizzed biennially (for as long as they survived) and also provided information about their medical and family histories and their socioeconomic background.

The people in the survey seem to have adjusted their personal life expectancies fairly appropriately for other information about themselves. People whose parents died younger were, quite properly, less optimistic about their own long-term survival. Current smokers discounted their futures by something like 6-8 years on average, which is in the right neighbourhood for a longtime smoker. For those respondents who were surveyed repeatedly, a major new diagnosis between survey waves did make them “correct” their life expectancies downward.

Perhaps the most interesting finding in the study is that English people displayed evidence of having information about their prospects over and above the other information given in the survey. Imagine neighbours Alfie and Barry who have the same age, health and genetic history, education, and socioeconomic status: for some reason, however, a glum Alfie tells the interviewer his chances of living another 10 years are poor, while Barry quite likes his odds.

The ELSA data suggest that ebullient Barry really is likely to outlast Alfie. Barry and Alfie know something relevant — probably no more complicated than “How they feel getting up in the morning” — that is not captured in a mere recitation of life facts.

Does all this mean that people are good at estimating their own life expectancies? Not so fast: what the study finds, and what other studies of its kind have found, is that there is a strong general pattern in people’s mortality estimates. When we are in our 50s and 60s, and perhaps really feeling the creep of decrepitude for the first time, we strongly underestimate our chances of living to advanced ages.

If you add up the ELSA numbers for 65-year-old men born in the 1940s, for example, you find that they collectively believe that only about two-thirds of them will reach age 75. We know from national life tables that this is preposterous: the real figure for that generation is 83 per cent. And women give just the same collective answer, even though women live longer, and the realworld survival rate for that generation is a whopping 89 per cent.

This “survival pessimism” exists for people under 70, no matter how you divide them up into subgroups, and so will influence key decisions about financial planning, possible annuity purchases, and pensions, at just the moment humans are making them. Then, as time goes by, the opposite problem appears: when people are in their 80s they begin to significantly overestimate how long they can expect to live.

This can have effects on economic well-being, too. Nobody, except perhaps their heirs, wants octogenarians to dine on cat food in an unheated shack and die with a fortune. Economists are sometimes accused of treating human beings as “homo economicus” — as super-rational, unfeeling automata who little resemble the frightened apes we really are. But no one captures our slightly absurd apelike features — as I suspect this study does — quite like an economist testing those “homo economicus” assumptions.